CN120563192A

CN120563192A - Commodity abnormality detection method and device and electronic equipment

Info

Publication number: CN120563192A
Application number: CN202510541975.3A
Authority: CN
Inventors: 王子林; 池昭波; 倪育丰; 赵琦; 赵锐娟
Original assignee: Taobao China Software Co Ltd
Current assignee: Taobao China Software Co Ltd
Priority date: 2025-04-27
Filing date: 2025-04-27
Publication date: 2025-08-29

Abstract

The embodiment of the application provides a commodity abnormality detection method and device and electronic equipment. The method for detecting the abnormality of the commodity comprises the steps of obtaining the commodity to be detected, determining the number of detection periods when the commodity to be detected is in an abnormal state continuously in the current detection period and a dynamic threshold value corresponding to the current detection period, and detecting the abnormality of the commodity to be detected based on the number of detection periods and the dynamic threshold value. According to the embodiment of the application, the commodities to be detected in the short-term abnormal state can be effectively screened, and as the commodities concerned by operators are the commodities in the long-term abnormal state, the quantity of the commodities concerned is effectively reduced through the operation, so that the platform operation is prevented from carrying out multiple invalid communication between the commodities concerned and merchants, the operation efficiency of the platform is improved, and the operation management cost is reduced.

Description

Commodity abnormality detection method and device and electronic equipment

Technical Field

The present application relates to the field of electronic commerce technologies, and in particular, to a method and an apparatus for detecting an abnormality of a commodity, and an electronic device.

Background

In the field of electronic commerce, abnormal states of commodities (e.g., states such as inventory type abnormality and price type abnormality) have a direct influence on commodity conversion rate. With the expansion of the platform scale, the types and the number of commodities are rapidly increased, and in order to ensure the real-time performance of the commodity abnormality detection operation, the existing commodity abnormality detection system generally adopts a higher detection frequency, which often leads to the generation of a large number of abnormal commodity results. When the platform operation receives the commodity abnormality alarm corresponding to the abnormal commodity result, the corresponding merchant is often informed immediately to process the abnormal commodity.

However, if the abnormal commodity is detected each time, the platform operation is triggered to notify the merchant, when the platform operation is required to simultaneously process a large number of abnormal commodities, the abnormal state of the commodity may be disappeared, namely, the normal state is restored, so that multiple invalid communications are performed between the platform operation and the merchant, and the maintenance and management cost of the platform operation is increased.

Disclosure of Invention

The embodiment of the application provides a commodity abnormality detection method, a commodity abnormality detection device and electronic equipment, which can avoid multiple invalid communication between platform operation and merchants to a certain extent, thereby reducing maintenance and management costs of the platform operation.

The embodiment of the invention provides a commodity abnormality detection method, which comprises the following steps:

acquiring a commodity to be detected, wherein the commodity to be detected is in an abnormal state in a current detection period;

determining the number of detection periods when the commodity to be detected is in an abnormal state continuously and a dynamic threshold corresponding to the current detection period of the commodity to be detected;

And carrying out anomaly detection on the commodity to be detected based on the detection period number and the dynamic threshold value.

The embodiment of the invention provides an abnormality detection device for commodities, which comprises:

The first acquisition module is used for acquiring the commodity to be detected, wherein the commodity to be detected is in an abnormal state in the current detection period;

The first determining module is used for determining the number of detection periods when the commodity to be detected is in an abnormal state continuously and a dynamic threshold value corresponding to the current detection period of the commodity to be detected;

And the first processing module is used for carrying out anomaly detection on the commodity to be detected based on the detection period number and the dynamic threshold value.

The embodiment of the invention provides electronic equipment, which comprises a memory and a processor, wherein the memory is used for storing one or more computer instructions, and the method in the first aspect is realized when the one or more computer instructions are executed by the processor.

An embodiment of the present invention provides a computer storage medium storing a computer program which, when executed by a computer, implements the method described in the first aspect.

Embodiments of the present invention provide a computer program product comprising a computer-readable storage medium storing computer instructions that, when executed by one or more processors, cause the one or more processors to perform the steps in the method described in the first aspect above.

According to the method, the device and the electronic equipment for detecting the abnormality of the commodity, the number of detection periods of the commodity to be detected which is continuously in the abnormality state and the dynamic threshold corresponding to the current detection period of the commodity to be detected are determined through the acquisition of the commodity to be detected, wherein the commodity to be detected is in the abnormality state in the current detection period, then the commodity to be detected in the abnormality state can be subjected to abnormality detection operation again based on the number of detection periods of the commodity to be detected which is continuously in the abnormality state and the dynamic threshold corresponding to the current detection period, so that the commodity to be detected in the abnormality state in a short period and the commodity to be detected in the abnormality state in a long period can be effectively identified.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:

fig. 1 is a schematic view of a scenario of a method for detecting an anomaly of a commodity according to an exemplary embodiment of the present application;

FIG. 2 is a schematic flow chart of a method for detecting anomalies in a commodity according to an exemplary embodiment of the present application;

FIG. 3 is a schematic flow chart of determining a dynamic threshold value corresponding to a current detection period of the commodity to be detected according to an exemplary embodiment of the present application;

FIG. 4 is a schematic flow chart of acquiring an article to be inspected according to an exemplary embodiment of the present application;

fig. 5 is a schematic diagram of detecting an abnormality of an alternative commodity in an alternative commodity set according to an embodiment of the present application by using a detection rule and commodity information of the alternative commodity;

FIG. 6 is a flowchart of another method for detecting anomalies in a commodity according to an exemplary embodiment of the present application;

FIG. 7 is a signaling interaction diagram of a method for detecting anomalies in a commodity according to an exemplary embodiment of the present application;

FIG. 8 is a schematic diagram of a system for detecting anomalies in a commodity according to an exemplary embodiment of the present application;

fig. 9 is a schematic structural diagram of an abnormality detection apparatus for a commodity according to an exemplary embodiment of the present application;

Fig. 10 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to specific embodiments of the present application and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

It should be noted that, in the case where the embodiment of the present application relates to user information, the user information (including, but not limited to, user equipment information, user personal information, etc.) and the data (including, but not limited to, data for analysis, stored data, presented data, etc.) related to the embodiment of the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data need to comply with the relevant laws and regulations and standards of the relevant country and region, and are provided with corresponding operation entries for the user to select authorization or rejection. In addition, the various models (including but not limited to language models or large models) to which the present application relates are compliant with relevant legal and standard regulations.

In addition, it should be noted that, in the case where the embodiment of the present application relates to a user interaction operation or trigger operation, the user interaction operation or trigger operation related to the embodiment of the present application includes, but is not limited to, an interaction operation in various manners, such as a touch operation, a gesture operation, a voice operation, a head movement operation, an eye movement operation, and the like, where the touch operation includes, but is not limited to, a click operation, a double click operation, a long press operation, a sliding operation, a pinch operation, or a mouse hovering operation. The sliding operation includes, but is not limited to, linear sliding, curved sliding, and the like.

Definition of terms:

LLM Large Language Model is a large-scale natural language processing model based on deep learning.

Commodity anomaly detection refers to identifying and processing abnormal states of commodities on an electronic commerce platform, such as backout, price fluctuation, activity expiration and the like.

Conversion efficiency, namely, the complete flow efficiency from commodity putting on shelf to final sales in the operation process of the electronic commerce platform.

Poisson distribution Poisson Distribution is a statistical probability distribution describing the number of events that occur in a fixed period of time. Poisson distribution is typically used to model the occurrence of sparse events, such as the number of events in anomaly detection.

In order to facilitate understanding of the method, the device and the electronic device for detecting abnormality of a commodity provided by the embodiments of the present application, related technologies are briefly described below:

In the application scenario of electronic commerce, the abnormal state of the commodity has a direct influence on the user experience and sales conversion efficiency. As the scale of platforms increases, the variety and number of goods increases rapidly, and systems for anomaly detection operations on goods face significant challenges:

(1) Contradiction between real-time performance and mass data

To ensure real-time performance, existing commodity anomaly detection systems typically employ relatively high detection frequencies, which often result in a large number of anomalous commodity results.

However, since the merchant usually repairs the abnormal state of the commodity in real time, the abnormal state of many commodities is rapidly released, so that the abnormal state of the commodity is outdated. The method makes it difficult for an operation team to accurately locate the true abnormal commodity to be concerned when processing a large amount of abnormal commodity data, thereby not only increasing the workload and reducing the efficiency and accuracy of the abnormal detection operation, but also failing to provide practical and effective support for the operation team.

(2) Challenges for large-scale commodity and multi-dimensional detection

Under the condition of numerous detection dimensions of commodity magnitude, an operation team is difficult to accurately sense the situation of the whole commodity, so that the comprehensiveness and accuracy of an operation decision can be influenced.

In order to solve the above technical problems, the embodiment of the present application provides a method, a device and an electronic device for detecting an abnormality of a commodity, and referring to fig. 1, an execution body of the method for detecting an abnormality of a commodity may be an abnormality detection device 200 of a commodity, where the abnormality detection device 200 of a commodity may be implemented as a local server or a cloud server. In the case where the abnormality detection apparatus 200 for a commodity is implemented as a cloud server, the abnormality detection method for the commodity may be executed in the cloud, where a plurality of computing nodes (cloud servers) may be deployed in the cloud, and each computing node has processing resources such as computation and storage. At the cloud, a service may be provided by multiple computing nodes, although one computing node may provide one or more services. The cloud may provide the service by providing a service interface to the outside, and the user invokes the service interface to use the corresponding service. The service interface includes a software development kit (Software Development Kit, abbreviated as SDK), an application program interface (Application Programming interface, abbreviated as API), and the like.

The abnormality detection device 200 of the commodity is communicatively connected to the client 100, where the client 100 is used for a user to apply to trigger an abnormality detection operation of the commodity, and the client 100 may be any computing device with a certain information interaction capability, and in specific implementation, the client 100 may be a mobile phone, a personal computer PC, a tablet computer, a set application program, or the like. Further, the basic structure of the client 100 may include at least one processor. The number of processors depends on the configuration and type of client. The client 100 may also include Memory that may be volatile, such as random access Memory (Random Access Memory, RAM) or non-volatile, such as Read-Only Memory (ROM), flash Memory, etc., or both. The memory typically stores an Operating System (OS), one or more application programs, program data, and the like. In addition to the processing unit and memory, the client 100 also includes some basic configurations, such as a network card chip, an IO bus, a display component, and some peripheral devices. Alternatively, some peripheral devices may include, for example, a keyboard, a mouse, a stylus, a printer, etc. Other peripheral devices are well known in the art and are not described in detail herein.

The commodity abnormality detection apparatus 200 is a device capable of performing an abnormality detection operation of a commodity in a network virtual environment, and generally is an apparatus for performing information planning and an abnormality detection operation of a commodity by using a network. The commodity abnormality detection apparatus 200 may be implemented as a commodity abnormality detection model for implementing commodity abnormality detection operations, and in physical implementation, the commodity abnormality detection apparatus 200 may be any device that can provide a computing service and can perform corresponding commodity abnormality detection operations, for example, may be a processor, a server, or the like. The commodity abnormality detection apparatus 200 mainly includes a processor, a hard disk, a memory, a system bus, and the like, and is similar to a general-purpose computer architecture.

In the present embodiment described above, the abnormality detection device 200 of the commodity and the client 100 are connected by a network, and the network connection may be a wireless or wired network connection. If the anomaly detection device 200 of the commodity is in communication connection with the client 100, the network system of the mobile network may be any of 2G (global system for mobile communications GSM), 2.5G (general packet radio service GPRS), 3G (wideband code division multiple access (WCDMA), time division synchronous code division multiple access (TD-SCDMA), 4G (long term evolution LTE), 4g+ (enhanced long term evolution lte+), worldwide interoperability for microwave access (WiMax), 5G, 6G, and the like.

In the embodiment of the present application, the client 100 is configured to be used by a user to generate or obtain an anomaly detection request for a commodity to be detected, where the number of commodities to be detected corresponding to the anomaly detection request may be one or more, and when the number of commodities to be detected is more, the multiple commodities to be detected may form a commodity set to be detected. In some examples, the abnormality detection request may be implemented through man-machine interaction, specifically, an interaction interface for performing interaction with the abnormality detection device 200 of the commodity may be displayed on the client 100, and the user may input an execution operation in the interaction interface, where the execution operation may be man-machine interaction or natural language voice interaction, so that the abnormality detection request for the commodity to be detected may be generated and obtained. In order to enable an abnormality detection operation for the commodity to be detected, the client 100 may transmit an abnormality detection request to the abnormality detection device 200 of the commodity.

The article anomaly detection device 200 is configured to obtain an anomaly detection request sent by the client 100, and then determine an article to be detected corresponding to the anomaly detection request, where the article to be detected is in an anomaly state in a current detection period, that is, it is indicated that an anomaly detection operation has been performed on the article to be detected in the current detection period, and a result of the detection operation is that the article to be detected is in an anomaly state. Because the number of the commodities in the abnormal state (namely, the commodities in the abnormal state in a short period) is large in a single detection period, in order to avoid invalid communication between platform operation and merchants for multiple times, repeated abnormal detection operation is required to be carried out on the commodities to be detected in the abnormal state in the current detection period, at this time, the number of the detection periods of the commodities to be detected in the abnormal state continuously and the dynamic threshold value corresponding to the commodities to be detected in the current detection period can be determined, wherein the dynamic threshold value corresponds to the detection period, namely, the dynamic threshold value can change along with the change of the detection period, and the dynamic threshold values corresponding to different detection periods can be different, so that the accuracy and the reliability of the abnormal detection operation on the commodities to be detected can be effectively improved when the commodities to be detected are subjected to abnormal detection based on the number of the detection periods and the dynamic threshold value.

In this embodiment, the anomaly detection operation can be performed again on the to-be-detected commodity in the anomaly state based on the number of detection periods in which the to-be-detected commodity is continuously in the anomaly state and the dynamic threshold corresponding to the current detection period, so that the to-be-detected commodity in the anomaly state in a short period and the to-be-detected commodity in the anomaly state in a long period are effectively identified and screened, and the commodity in the anomaly state in a long period is the commodity to be focused, and then the operation management and the maintenance operation of the commodity to be focused can be performed based on the anomaly detection result, so that not only is the ineffective communication performed between the platform operation based on the to-be-detected commodity in the anomaly state in a short period and the merchant for many times, but also the operation maintenance and the management cost of the platform is reduced, and the practicability of the method is further improved.

The following describes in detail the technical solutions provided by the embodiments of the present application with reference to the accompanying drawings.

Fig. 2 is a flow chart of a method for detecting an abnormality of a commodity according to an exemplary embodiment of the present application, and referring to fig. 2, the present embodiment provides a method for detecting an abnormality of a commodity, an execution subject of the method is an abnormality detection device of a commodity, the abnormality detection device of a commodity may be implemented as software or a combination of software and hardware, when the abnormality detection device of a commodity is implemented as hardware, it may be specifically various electronic devices capable of implementing an abnormality detection operation of a commodity, including, but not limited to, a personal computer, a server, etc., and when the abnormality detection device of a commodity is implemented as software, it may be installed in the electronic device exemplified above. Specifically, the method for detecting abnormality of a commodity provided in this embodiment may include:

step S201, acquiring a commodity to be detected, wherein the commodity to be detected is in an abnormal state in the current detection period.

Step S202, determining the number of detection periods when the commodity to be detected is in an abnormal state continuously and a dynamic threshold corresponding to the current detection period of the commodity to be detected.

And step S203, carrying out anomaly detection on the commodity to be detected based on the number of detection periods and the dynamic threshold value.

The specific implementation manner and implementation principle of each step are described in detail below:

The article to be detected may be an article that needs to perform an abnormal detection operation, where the abnormal operation may be implemented as a periodic detection operation according to a preset detection period, and it should be noted that the article to be detected is not an ordinary article that needs to perform an abnormal detection operation, but an article that is in an abnormal state in a current detection period, where the abnormal state may include at least one of an article to be detected being in an out-of-stock state in the current detection period, an article to be detected being in a fluctuating state in a price of the current detection period, an article to be detected having an activity expired state in the current detection period, and the like.

For example, for the same commodity, when the number of detection periods includes 3, the detection period T1, the detection period T2, the detection period T3 are included, and when the detection period T3 is the current detection period, the history detection period includes the detection period T1 and the detection period T2. When the detection state of the commodity in the detection period T2 is an abnormal state and the detection state in the detection period T3 is a normal state, the commodity will not be the commodity to be detected in the embodiment of the present application. Correspondingly, when the detection state of the commodity in the detection period T2 is a normal state and the detection state in the detection period T3 is an abnormal state, the commodity may be the commodity to be detected in the embodiment of the present application.

In some examples, the commodity to be detected can be obtained through a client, and the obtaining of the commodity to be detected can comprise determining the client which is in communication connection with an abnormality detection device of the commodity, wherein the client comprises the commodity to be detected which is in an abnormality state in the current detection period, and the commodity to be detected is obtained actively or passively based on the client, so that the accuracy and the reliability of obtaining the commodity to be detected are ensured to a certain extent.

In other examples, the commodity to be detected can be obtained through the client, and can be determined through analysis processing on the commodity state of the commodity in the current detection period, and the obtaining of the commodity to be detected can comprise obtaining a plurality of alternative commodities, determining the preliminary detection states corresponding to the plurality of alternative commodities in the current detection period, determining the alternative commodity as the commodity to be detected when the preliminary detection state of the alternative commodity is an abnormal state, and determining the alternative commodity as not the commodity to be detected when the preliminary detection state of the alternative commodity is a normal state, so that the accuracy and the reliability of obtaining the commodity to be detected are ensured.

Because the number of the commodities with abnormal states in a single detection period is often more when the preliminary abnormal detection operation is performed on the mass commodities existing in the electronic commerce platform, in order to avoid the platform operation from performing multiple invalid communications between the commodities with abnormal states and merchants, after the commodities to be detected are acquired, the repeated abnormal detection operation can be performed on the commodities to be detected, so that whether the commodities to be detected are operators or the commodities which the operation platform really needs to pay attention to can be identified.

In order to realize the repeated abnormal detection operation of the commodity to be detected, after the commodity to be detected is obtained, the detection period number of the commodity to be detected in an abnormal state continuously can be determined, wherein the detection period number can be determined through the historical detection result of the commodity to be detected in the historical detection period and the current detection result in the current detection period, and at the moment, the determination of the detection period number of the commodity to be detected in the abnormal state continuously can comprise the steps of obtaining the historical detection result corresponding to the commodity to be detected in the historical detection period and the current detection result corresponding to the current detection period, and determining the detection period number of the commodity to be detected in the abnormal state continuously based on the historical detection result and the current detection result.

Specifically, the current relevant data of the current detection period and the historical relevant data of the historical detection period can be obtained for the commodity to be detected, and then the abnormality detection operation can be performed on the commodity to be detected in the historical detection period and the commodity to be detected in the current detection period based on the current relevant data and the historical relevant data respectively. In some examples, the abnormality detection operation may be implemented by a state detection model for implementing the abnormality detection operation of the commodity, so that a history detection result and a current detection result may be obtained.

And then analyzing and processing the historical detection result and the current detection result to obtain the number of detection periods of the commodity to be detected in the abnormal state continuously, specifically, when the current detection result is the abnormal detection result, identifying whether the historical detection result adjacent to the current abnormal detection result is the abnormal detection result, if so, continuously counting whether the historical detection result adjacent to the historical abnormal detection result is the abnormal detection result, if so, executing the operation until the historical detection result is not the abnormal detection result, and determining the number of detection periods in the abnormal state continuously based on the current abnormal detection result and the number of continuous historical abnormal detection results.

The state detection model in the embodiment of the application can be a Language Model (LM) based on artificial intelligence or a multi-Mode model (Multimodal Model MM) based on artificial intelligence, and the embodiment of the application does not limit the number of model parameters supported by the model so as to meet the actual requirements.

For example, when the current detection result corresponding to the current detection period is in an abnormal state, whether the history detection result corresponding to the history detection period is in an abnormal state may be queried forward from the current detection result corresponding to the current detection period, and when the history detection result is in an abnormal state, the number of history detection periods continuously in an abnormal state may be counted, and when the number of continuous history detection periods in an abnormal state is 2, since the current detection result is in an abnormal state, it may be determined that the number of detection periods continuously in an abnormal state of the commodity to be detected is 3. When the historical detection result corresponding to the historical detection period adjacent to the current detection period is in a normal state, the number of detection periods of the commodity to be detected in an abnormal state continuously can be determined to be 1, so that the accuracy and reliability of determining the number of detection periods of the commodity to be detected in the abnormal state continuously are effectively ensured.

Because the anomaly detection operation of the commodity to be detected again needs to determine not only the number of detection periods when the commodity to be detected is continuously in an anomaly state, but also the dynamic threshold value corresponding to the current detection period of the commodity to be detected, in some examples, the dynamic threshold value can be determined through a preconfigured mapping relation corresponding to the detection period, and at this time, determining the dynamic threshold value corresponding to the current detection period of the commodity to be detected can comprise obtaining the preconfigured mapping relation between each detection period and each dynamic threshold value, and determining the dynamic threshold value corresponding to the current detection period based on the mapping relation and the identity of the current detection period, so that the accuracy and reliability of determining the dynamic threshold value are improved to a certain extent.

After the detection period number of the commodities to be detected in the abnormal state continuously and the dynamic threshold value corresponding to the current detection period of the commodities to be detected are obtained, the operations of detecting the abnormality of the commodities to be detected can be carried out based on the detection period number and the dynamic threshold value, so that whether the commodities to be detected are the commodities to be detected in the abnormal state in a short period can be effectively identified.

Further, when the number of the commodities to be detected is plural, an abnormality detection operation may be performed for each commodity to be detected to identify whether any commodity to be detected is a commodity to be detected in an abnormal state for a short period, so that for the plural commodities to be detected, the commodity to be detected in an abnormal state for a short period and the commodity to be detected in an abnormal state for a long period included in the plural commodities to be detected can be effectively screened.

In some examples, the anomaly detection operation may be implemented by a pre-trained anomaly detection model, where performing anomaly detection on the commodity to be detected based on the number of detection periods and the dynamic threshold may include obtaining the pre-trained anomaly detection model, inputting the number of detection periods, the dynamic threshold, and commodity data corresponding to the commodity to be detected into the anomaly detection model to perform the anomaly detection operation, and obtaining an anomaly detection result. The abnormality detection result may include a first result for identifying that the commodity to be detected is in an abnormal state, or the abnormality detection result may include a second result for identifying that the commodity to be detected is in a normal state.

In other examples, the anomaly detection operation may be implemented by not only analyzing the number of detection periods, the dynamic threshold value, and the commodity data corresponding to the commodity to be detected through the anomaly detection model, but also comparing the number of detection periods with the dynamic threshold value, where the anomaly detection of the commodity to be detected based on the number of detection periods and the dynamic threshold value may include determining that the commodity to be detected is an anomalous commodity if the number of detection periods is greater than or equal to the dynamic threshold value, or determining that the commodity to be detected is a non-anomalous commodity if the number of detection periods is less than the dynamic threshold value.

The dynamic threshold value can be used for identifying the detection cycle lower limit value of the to-be-detected commodity in an abnormal state, after the detection cycle number and the dynamic threshold value are obtained, the detection cycle number can be compared with the dynamic threshold value in an analysis mode, and under the condition that the detection cycle number is larger than or equal to the dynamic threshold value, the current to-be-detected commodity can be determined to be the commodity in the abnormal state after meeting the detection condition in the abnormal state, and under the condition that the detection cycle number is smaller than the dynamic threshold value, the current to-be-detected commodity can be determined to be the commodity in the non-abnormal state after not meeting the detection condition in the abnormal state, so that the to-be-detected commodity can be accurately detected in the abnormal state.

In addition, after determining that the commodity to be detected is an abnormal commodity, the method in the embodiment can further comprise generating a detection result and an abnormal prompt message for identifying that the commodity to be detected is an abnormal commodity and outputting the detection result and the abnormal prompt message.

Specifically, after determining that the commodity to be detected is an abnormal commodity, a detection result for identifying that the commodity to be detected is an abnormal commodity may be generated, and abnormal prompt information corresponding to the detection result may be generated, where the detection result and the abnormal prompt information may be generated through a detection analysis model trained in advance.

After the detection result and the abnormal prompt information are obtained, the detection result and the abnormal prompt information can be associated and output, the detection result and the abnormal prompt information can be specifically output and displayed through a display device, or the detection result and the abnormal prompt information can be sent to operators in a communication message mode (such as a mail mode and an instant messaging software mode), so that the operators can quickly know the abnormal state of the commodity to be detected through the output detection result and the abnormal prompt information, and can flexibly adjust and configure the operation of the commodity to be detected, and further the quality and effect of the operation of the commodity to be detected can be improved to a certain extent, and meanwhile, the commodity conversion efficiency of the commodity to be detected can also be improved.

Further, after the to-be-detected commodity is detected abnormally based on the number of detection periods and the dynamic threshold, operation suggestion information corresponding to the to-be-detected commodity can be generated, so that an adjustment direction of operation can be provided for operators, and improvement of operation quality and effect of the to-be-detected commodity can be facilitated.

Specifically, after the to-be-detected commodity is subjected to abnormal detection based on the number of detection periods and the dynamic threshold value, an abnormal detection result obtained through abnormal detection operation can be analyzed and processed to generate operation suggestion information of the to-be-detected commodity, wherein the operation suggestion information can be determined through a pre-trained commodity analysis model, namely, the abnormal detection result and the to-be-detected commodity are input into the commodity analysis model for analysis and processing, and the operation suggestion information corresponding to the to-be-detected commodity, which is output by the commodity analysis model, is obtained. In some examples, the operational advice information may include at least one of content display advice corresponding to the item to be detected, advertising advice corresponding to the item to be detected, operational channel advice corresponding to the item to be detected, and the like.

After the operation suggestion information is obtained, the abnormal detection result and the operation suggestion information can be output in an associated manner, and the abnormal detection result and the operation suggestion information can be output and displayed through a display device or can be sent to an operator in a communication message mode, so that the operator can quickly know the state of the commodity to be detected through the output abnormal detection result and the operation suggestion information, and can flexibly adjust and configure the operation of the commodity to be detected, and further the quality and effect of the operation of the commodity to be detected can be improved to a certain extent.

According to the method for detecting the abnormality of the commodity, the number of detection periods of the commodity to be detected in the abnormal state continuously and the dynamic threshold corresponding to the current detection period of the commodity to be detected are determined by acquiring the commodity to be detected, wherein the commodity to be detected is in the abnormal state in the current detection period, then the commodity to be detected in the abnormal state can be subjected to the abnormality detection operation again based on the number of detection periods of the commodity to be detected in the abnormal state continuously and the dynamic threshold corresponding to the current detection period, so that the commodity to be detected in the abnormal state in the short period and the commodity to be detected in the abnormal state in the long period can be effectively identified.

Fig. 3 is a schematic flow chart of determining a dynamic threshold value corresponding to a current detection period of an article to be detected according to an exemplary embodiment of the present application, and based on the foregoing embodiment, referring to fig. 3, the dynamic threshold value may be determined not only by a preconfigured mapping relationship corresponding to the detection period, but also by analyzing a set of detected articles to which the article to be detected belongs, where determining the dynamic threshold value corresponding to the current detection period of the article to be detected may include:

Step S301, a detection commodity set to which a commodity to be detected belongs is obtained.

The method can be realized not only as a single commodity, but also as any one of the detected commodity sets, so that when the abnormality detection operation of the commodity is performed on the commodity to be detected, the abnormality detection operation can be performed not only on a certain commodity to be detected, but also on a plurality of commodities to be detected in a certain detected commodity set.

When the commodity to be detected is realized as any one of the detection commodity sets, the dynamic threshold is related to the detection period and the detection commodity set, so that the detection commodity set to which the commodity to be detected belongs can be acquired first in order to accurately determine the dynamic threshold corresponding to the commodity to be detected in the current detection period. In some examples, the detection commodity set can be determined through a preset mapping relation corresponding to the commodity to be detected, and at this time, acquiring the detection commodity set to which the commodity to be detected belongs can comprise acquiring a commodity identifier of the commodity to be detected, and determining the detection commodity set to which the commodity to be detected belongs by utilizing the preset mapping relation between the commodity identifier and the set identifier of the detection commodity set and the commodity identifier, so that the accuracy and the reliability of acquiring the detection commodity set are effectively ensured.

Step S302, determining the average abnormal occurrence rate of the detected commodity set in a preset time period, wherein the preset time period is greater than or equal to at least two detection periods.

Since the dynamic threshold is used for performing the abnormality detection operation on the commodity to be detected, which is often related to the average abnormality occurrence rate corresponding to the detected commodity set, in order to accurately determine the dynamic threshold corresponding to the commodity to be detected in the current detection period, after the detected commodity set to which the commodity to be detected belongs is obtained, the average abnormality occurrence rate corresponding to the detected commodity set in a preset time period can be determined, wherein the preset time period includes a plurality of detection periods.

In some examples, the average anomaly occurrence rate may be determined by a pre-trained anomaly detection model, where determining the average anomaly occurrence rate corresponding to the detected commodity set in the preset time period may include obtaining the pre-trained anomaly detection model, determining detected commodity data corresponding to each commodity to be detected in the detected commodity set in the preset time period, where the detected commodity data may include at least one of a commodity identification of the commodity to be detected, commodity operation information of the commodity to be detected, commodity name of the commodity to be detected, brand information of the commodity to be detected, description information of the commodity to be detected, inventory and pricing information of the commodity to be detected, and the like, and then inputting the detected commodity data corresponding to each commodity to be detected into the anomaly detection model for processing, so as to obtain the average anomaly occurrence rate output by the anomaly detection model, thereby ensuring accuracy and reliability of determining the average anomaly occurrence rate to a certain extent.

In other examples, the average anomaly occurrence rate may be determined not only by a pre-trained anomaly detection model, but also based on a total anomaly number corresponding to the detected commodity in the detected commodity set and a preset time period, where determining the average anomaly occurrence rate corresponding to the detected commodity set in the preset time period may include obtaining the total anomaly number corresponding to the detected commodity in the detected commodity set, determining a number of detection periods included in the preset time period, and determining the average anomaly occurrence rate based on the total anomaly number and the number of detection periods.

Because one detected commodity set often includes a plurality of detected commodities, different detected commodities can correspond to different state information within a preset time period, and the average abnormal occurrence rate is related to the abnormal state corresponding to the detected commodity in the detected commodity set, in order to accurately determine the average abnormal occurrence rate corresponding to the detected commodity set within the preset time period, the total abnormal times corresponding to the detected commodity in the detected commodity set can be acquired first, and the total abnormal times can be acquired by summarizing the abnormal times corresponding to each detected commodity in the detected commodity set.

Since the average abnormality occurrence rate is related not only to the total number of abnormalities corresponding to the detected products in the detected product set but also to the preset time period, for the preset time period, the preset time period may be subjected to analysis processing to determine the number of detection cycles included in the preset time period. In some examples, the number of detection periods may be determined by a preset time period and a detection period, where determining the number of detection periods included in the preset time period may include obtaining a period duration corresponding to the detection period, and dividing the preset time period based on the period duration to obtain the number of detection periods included in the preset time period, so that accuracy and reliability in determining the number of detection periods are ensured to a certain extent.

After the total number of anomalies and the number of detection cycles are obtained, the total number of anomalies and the detection cycles may be analyzed to determine an average anomaly occurrence rate. In some examples, determining the average anomaly occurrence rate based on the total anomaly number and the number of detection cycles may include determining a ratio of the total anomaly number to the number of detection cycles as the average anomaly occurrence rate. For example, when the Total number of anomalies is total_ Anomalies and the number of detection cycles is total_ Periods, the average anomaly occurrence rate λ can be determined to beThus, the accuracy and reliability of determining the average anomaly occurrence rate lambda are effectively ensured.

In other examples, the average anomaly occurrence rate may be determined not only by a ratio of the total anomaly number to the number of detection periods, but also by a pre-trained occurrence rate calculation model, where determining the average anomaly occurrence rate based on the total anomaly number and the number of detection periods may include obtaining an occurrence rate determination model, which may be a pre-trained network model for calculating the average anomaly occurrence rate or may be a large language model capable of calculating the average anomaly occurrence rate, and then processing the total anomaly number and the number of detection periods using the occurrence rate determination model, so as to determine the average anomaly occurrence rate, and specifically, inputting the total anomaly number and the number of detection periods into the occurrence rate determination model for analysis, so as to obtain the average anomaly occurrence rate output by the occurrence rate determination model, thereby also ensuring the accuracy and reliability of determining the average anomaly occurrence rate.

Step S303, determining a dynamic threshold based on the average abnormal occurrence rate and the preset confidence.

After the average abnormal occurrence rate is obtained, the average abnormal occurrence rate and a preset confidence coefficient can be analyzed and processed to determine a dynamic threshold, wherein the preset confidence coefficient is a pre-configured parameter for describing the reliability degree of the dynamic threshold, so that the accuracy and the reliability of determining the dynamic threshold are ensured to a certain extent.

In some examples, the dynamic threshold may be determined by a pre-trained threshold calculation model, where determining the dynamic threshold based on the average anomaly occurrence and the preset confidence level may include obtaining the pre-trained threshold calculation model, inputting the average anomaly occurrence and the preset execution level into the threshold calculation model for analysis, and obtaining the dynamic threshold output by the threshold calculation model, so that accuracy and reliability of determining the dynamic threshold are ensured to a certain extent.

In other examples, the dynamic threshold may be determined not only by a pre-trained threshold calculation model, but also by a cumulative distribution function of the poisson distribution, where determining the dynamic threshold based on the average anomaly occurrence rate and the preset confidence level may include obtaining the cumulative distribution function of the poisson distribution, and processing the average anomaly occurrence rate and the preset confidence level using the cumulative distribution function to obtain the dynamic threshold output by the cumulative distribution function.

On the premise of not losing generality, the condition that the commodity to be detected changes into an abnormal state to be a random process at a certain moment and is irrelevant to the state at the last moment can be assumed, so that the state at the next moment meets poisson distribution, and a dynamic threshold value can be determined by using a poisson distribution model. In particular, a cumulative distribution function of the poisson distribution may be obtained first, which Cumulative Distribution Function (CDF) gives the probability that the random variable (X) is less than or equal to a certain value (k). And then the average abnormal occurrence rate and the preset confidence coefficient can be processed by utilizing the cumulative distribution function, so that the dynamic threshold value output by the cumulative distribution function can be obtained stably.

Further, the average abnormal occurrence rate and the preset confidence coefficient are processed by using the cumulative distribution function, the dynamic threshold value output by the cumulative distribution function can be obtained by obtaining a Probability Mass Function (PMF) of Poisson distribution, the Probability Mass Function (PMF) gives the probability of the occurrence (k) times of the event in a certain interval, and then the average abnormal occurrence rate and the preset confidence coefficient are processed by using the probability mass function and the cumulative distribution function, so that the dynamic threshold value can be obtained stably, and the accuracy and the reliability of determining the dynamic threshold value are ensured.

In the embodiment, the average abnormal occurrence rate corresponding to the detected commodity set in the preset time period is determined by acquiring the detected commodity set to which the commodity to be detected belongs, and then the dynamic threshold is determined based on the average abnormal occurrence rate and the preset confidence, so that the dynamic threshold is flexibly and stably determined, and the flexibility and reliability of the abnormality detection operation on the commodity are improved.

Fig. 4 is a schematic flow chart of acquiring a commodity to be detected according to an exemplary embodiment of the present application, and on the basis of any one of the foregoing embodiments, referring to fig. 4, the commodity to be detected may be determined not only by a client, but also by performing a preliminary detection operation on an alternative commodity set, where acquiring the commodity to be detected may include:

Step S401, acquiring an alternative commodity set, wherein the alternative commodity set comprises a plurality of alternative commodities, and commodity information corresponding to the alternative commodities.

The to-be-detected commodity can be any one of a plurality of alternative commodities, and in order to accurately obtain the to-be-detected commodity needing to perform the abnormality detection operation again, an alternative commodity set can be obtained first, wherein the alternative commodity set comprises a plurality of alternative commodities. In order to further improve the accuracy and reliability of acquiring the commodity to be inspected, the commodity information may correspond to the commodity candidates in the commodity candidate set, where the commodity information may include at least one of a commodity attribute (including at least one of a commodity name, a commodity number, a commodity brand, and a commodity category), commodity description information, inventory and pricing of the commodity, picture information of the commodity, logistics information of the commodity, a transaction record of the commodity, and the like.

In some examples, the set of alternative merchandise may be obtained by a client, wherein the set of alternative merchandise may be stored in the client, at which point the set of alternative merchandise may be actively or passively obtained by the client.

In other examples, the candidate commodity set can be obtained not only through the client, but also based on preset commodity screening information, and the obtaining of the candidate commodity set can comprise obtaining commodity screening information, identifying identity information corresponding to at least one candidate commodity which accords with the commodity screening information in a preset commodity database, and carrying out completion operation of the commodity information based on the identity information corresponding to the at least one candidate commodity to obtain the candidate commodity set.

Specifically, the commodity screening information may refer to information or conditions for performing a preliminary screening operation on the candidate commodities in the candidate commodity set, and in some examples, the commodity screening information may be obtained through a man-machine interaction operation or a client, where the commodity screening information may include at least one of a commodity screening rule and a commodity screening keyword. The commodity screening rule can comprise at least one of whether commodity attributes are matched with a delivery channel or not and whether the punishment state is normal, and the punishment state can comprise at least one of punishment state of commodities, punishment state of shops where the commodities are located and the like.

After acquiring the commodity screening information, identifying the identity information corresponding to each of at least one candidate commodity according with the commodity screening information in a preset commodity database, wherein the preset commodity database can be a database comprising commodity data corresponding to a large number of commodities, and in order to accurately acquire the candidate commodity set, screening the commodities in the commodity database by using the acquired commodity screening information, so that the identity information corresponding to each of the at least one candidate commodity can be acquired.

In order to accurately acquire the candidate commodity set, the candidate commodity information can be complemented based on the identity information corresponding to at least one candidate commodity, specifically, the information search can be performed in a preset database based on the identity information corresponding to at least one candidate commodity to acquire the commodity information, and then the completion operation of the commodity information is completed, so that the at least one candidate commodity corresponding to the commodity information can be acquired, and the candidate commodity set formed by the at least one candidate commodity and the commodity information corresponding to each candidate commodity can be acquired, so that the accuracy and the reliability of acquiring the candidate commodity set are ensured to a certain extent.

Step S402, determining a detection rule for performing preliminary detection operation on the candidate commodity set.

Since the candidate commodity in the candidate commodity set may be in a normal state or an abnormal state, and the commodity to be detected refers to the commodity in the abnormal state in the current detection period, in order to accurately acquire the commodity to be detected, a detection rule for performing a preliminary detection operation on the candidate commodity set may be determined. In some examples, the detection rules may include at least one of whether the price of the merchandise is compatible with the delivery channel, whether the merchandise corresponds to preset equity information (e.g., preset merchandise sales promotion activity, preset merchandise discount activity, etc.), whether there is inventory in a preset area, and whether the inventory value of the merchandise is greater than a preset threshold.

Specifically, the determining manner of the detection rule in this embodiment is similar to the obtaining manner of the commodity to be detected in the above embodiment, and reference may be made to the above description specifically, and details are not repeated here.

And S403, performing anomaly detection on the candidate commodity in the candidate commodity set by utilizing the detection rule and commodity information of the candidate commodity to obtain a detection result of the candidate commodity.

After the detection rule and the commodity information of the candidate commodity are acquired, the detection rule and the commodity information of the candidate commodity can be utilized to perform abnormal detection operation on the candidate commodity in the candidate commodity set, so that the detection result of the candidate commodity can be obtained. The detection result may be a first result for identifying that the candidate commodity is in a normal state, or the detection result may be a second result for identifying that the candidate commodity is in an abnormal state.

The detection rule may include a plurality of detection sub-rules, and when the detection rule and the commodity information of the candidate commodity are used to perform an abnormal detection operation on the candidate commodity in the candidate commodity set, the abnormal detection operation may be sequentially performed on the candidate commodity according to the detection sub-rules and the data of the candidate commodity, so that a detection result of the candidate commodity may be obtained. For example, referring to fig. 5, one of the detection sub-rules may be pulled first, and data of an alternative commodity may be obtained, and then an engine may be used to perform an anomaly detection operation based on the data of the alternative commodity and the detection sub-rule, so as to obtain a sub-detection result corresponding to the detection sub-rule. Whether the sub detection result identifies whether the candidate commodity passes the abnormality detection operation or not is identified, if the candidate commodity does not pass the abnormality detection operation, the candidate commodity can be determined to be in an abnormal state, a corresponding abnormality record can be generated based on the candidate commodity, and then the corresponding abnormality detection result can be generated based on the abnormality record.

Correspondingly, when the candidate commodity is determined to pass through the abnormal detection operation corresponding to one detection sub-rule, whether all detection rules are detected is identified, and under the condition that the abnormal detection operation is not carried out on the candidate commodity by using all detection rules, the next detection sub-rule can be pulled out, and the abnormal detection operation is carried out on the candidate commodity by using the next detection sub-rule and the data of the candidate commodity, so that a detection sub-result corresponding to the next detection sub-rule is obtained. By the time the abnormality detection operation is performed on the candidate commodity by using all detection rules, an abnormality detection result corresponding to the candidate commodity can be generated based on the detection sub-results corresponding to all detection sub-rules, so that the accuracy and the reliability of determining the abnormality detection result are effectively ensured.

It should be noted that, for the detection rule including a plurality of detection sub-rules, not only one engine may be utilized to sequentially perform the abnormality detection operation for the candidate commodity based on one detection sub-rule for the plurality of detection sub-rules, respectively, or a plurality of engines may be utilized to synchronously perform the abnormality detection operation for the candidate commodity using a different plurality of detection sub-rules, so that the quality and efficiency of the abnormality detection operation are ensured to some extent.

And step S404, determining the commodity candidate corresponding to the detection result as the commodity to be detected under the condition that the detection result is used for identifying the commodity candidate as the abnormal state.

When the detection result is used for identifying that the alternative commodity is in a normal state, the alternative commodity is in a normal state in the current detection period, so that the alternative commodity corresponding to the detection result can be determined as not being the commodity to be detected, and when the detection result is used for identifying that the alternative commodity is in an abnormal state, the alternative commodity is in an abnormal state in the current detection period, so that the alternative commodity corresponding to the detection result can be determined as the commodity to be detected.

In the embodiment, the detection rule for performing the preliminary detection operation on the candidate commodity set is determined by acquiring the candidate commodity set, then the detection rule and the commodity information of the candidate commodity are utilized to perform the abnormal detection on the candidate commodity in the candidate commodity set, the detection result of the candidate commodity is obtained, and in the case that the detection result is used for identifying that the candidate commodity is in an abnormal state, the candidate commodity corresponding to the detection result can be determined as the commodity to be detected, so that the flexible reliability of acquiring the commodity to be detected is effectively ensured.

Fig. 6 is a schematic flow chart of another method for detecting an abnormality of a commodity according to an exemplary embodiment of the present application, and, based on any one of the foregoing embodiments, referring to fig. 6, after detecting an abnormality of a commodity to be detected based on the number of detection periods and a dynamic threshold, an abnormal commodity trend corresponding to a detected commodity set may be generated, where the method in this embodiment may further include:

step S601, obtaining a detection commodity set corresponding to the commodity to be detected.

In order to provide comprehensive visual angle information and related state trends about the detected commodity and help operators make a more intelligent decision in a complex environment, after the to-be-detected commodity is abnormally detected based on the number of detection periods and the dynamic threshold, in order to accurately determine the abnormal commodity trend of the to-be-detected commodity set corresponding to the to-be-detected commodity, the to-be-detected commodity set corresponding to the to-be-detected commodity may be acquired first.

The method for acquiring the detected commodity set in this embodiment is similar to the method for acquiring the detected commodity set in the above embodiment, and specific reference may be made to the descriptions in the above embodiment, which are not repeated here.

Step S602, determining a current detection result of the detected commodity set in a current detection period and a historical detection result in a historical detection period.

After the detected commodity set is acquired, the detected commodity set may be analyzed to determine a current detection result of the detected commodity set in a current detection period and a historical detection result in a historical detection period. The method comprises the steps of obtaining a pre-trained abnormality detection model, carrying out abnormality detection operation on each detection commodity in the detection commodity set in the current detection period by using the abnormality detection model to obtain commodity detection results corresponding to each detection commodity, and then carrying out statistics on commodity detection results corresponding to all detection commodities, so that the current detection result of the detection commodity set in the current detection period can be obtained, and the accuracy and the reliability of determining the current detection result are ensured to a certain extent.

In addition, the specific determination manner of the history detection result of the detected commodity set in the history detection period is similar to the specific determination manner of the current detection result in the current detection period of the detected commodity set, and the specific reference is made to the above statement content, and the description is omitted here. It should be noted that each of the above-described history detection periods corresponds to one history detection result, and when the number of history detection periods is plural, the number of obtained history detection results is plural.

Step S603, generating an abnormal commodity trend corresponding to the detected commodity set based on the current detection result and the historical detection result.

After the current detection result and the historical detection result are obtained, the current detection result and the historical detection result can be analyzed, and particularly, the current detection result and the historical detection result can be analyzed, so that an abnormal commodity trend corresponding to the detected commodity set can be generated. The number of the abnormal commodity trends can be multiple, different abnormal commodity trends can correspond to different parameter dimensions, and the different parameter dimensions can comprise at least one of the dimension of commodity stock, the dimension of commodity abnormal rate, the dimension of commodity price and the like.

In some examples, the abnormal commodity trend can be generated through a pre-trained trend generating model, and generating the abnormal commodity trend corresponding to the detected commodity set based on the current detection result and the historical detection result can comprise the steps of acquiring the pre-trained trend generating model, inputting the current detection result and the historical detection result into the trend generating model for trend generating operation, and acquiring the abnormal commodity trend corresponding to the detected commodity set output by the trend generating model, so that the stability and the reliability of generating the abnormal commodity trend are effectively ensured.

Further, after generating the abnormal commodity trend corresponding to the detected commodity set, operation suggestion information corresponding to the detected commodity set may be generated, and at this time, the method in this embodiment may further include generating operation suggestion information corresponding to the detected commodity set based on the abnormal commodity trend, and performing association output on the abnormal commodity trend and the operation suggestion information.

Specifically, since different abnormal commodity trends can represent that the detected commodity set has different states, in order to enable a user to timely and quickly learn the state of the detected commodity set, after generating an abnormal commodity trend corresponding to the detected commodity set, the abnormal commodity trend can be analyzed and processed, so that operation suggestion information corresponding to the detected commodity set can be generated. The operation advice information is advice information related to commodity operation, and may include at least one of advice information related to commodity inventory, advice information related to the number of commodity offerings, advice information related to commodity popularization activities, and the like.

In some examples, the operation advice information may be determined by a pre-trained trend analysis model, and generating the operation advice information corresponding to the detected commodity set based on the abnormal commodity trend may include obtaining the pre-trained trend analysis model, and analyzing the abnormal commodity trend by using the trend analysis model to obtain the operation advice information corresponding to the detected commodity set output by the trend analysis model, so that the accuracy and reliability of the generation of the operation advice information are ensured to a certain extent.

In the embodiment, the current detection result of the detected commodity set in the current detection period and the historical detection result of the detected commodity set in the historical detection period are determined by acquiring the detected commodity set corresponding to the commodity to be detected, and then the abnormal commodity trend corresponding to the detected commodity set is generated based on the current detection result and the historical detection result, so that the accuracy and the reliability of generating the abnormal commodity trend are effectively ensured.

In a specific application, referring to fig. 7 to 8, this application embodiment provides a method for detecting an abnormality of a commodity, where an execution subject of the method is an abnormality detection system, and the abnormality detection system may use an abnormal commodity extraction model generated by a mathematical model to filter short-term and invalid abnormal commodities, and may combine with a large language model to generate a prediction trend, and then may perform a commodity operation based on the prediction trend, so that a more accurate management support operation can be provided for an operation team, and further quality and efficiency of the commodity operation can be improved to a certain extent. Specifically, the abnormality detection method implemented based on the abnormality detection system may include the steps of:

and 1, acquiring the commodity to be detected.

The main link for anomaly detection can acquire the commodity to be detected through the commodity acquisition service for detection, specifically can acquire the commodity to be detected according to a preset timing detection task, and the commodity acquisition service for detection can acquire the ID information of the commodity to be detected from a commodity library according to a preset commodity screening rule and a commodity screening keyword, and the timing detection task can trigger 1 commodity acquisition operation to be detected for every 1 hour, so that the periodic anomaly detection operation of the commodity can be realized.

In some examples, the merchandise screening rules may be screening rules of merchandise attribute dimensions, and specifically may include at least one of whether a picture/title matches a delivery channel and whether a penalty status is normal. For example, the screening rule may be whether the image/title matches the spring sales promotion channel, so that the merchandise to be detected related to the spring sales promotion channel can be screened out by the screening rule.

In addition, the commodity screening keywords can comprise the commodity characteristic information such as the retrieval tag information, the retrieval keywords or the retrieval activities input by the user, so that the commodity to be detected meeting the user requirements can be obtained through the commodity screening keywords. And the timing detection task can be realized as an offline task, so that the commodities to be detected meeting the commodity screening rule can be screened out from the full data corresponding to the electronic commerce platform through the offline task.

And 2, carrying out completion operation on the commodity to be detected.

Because the commodity to be detected only corresponds to the ID information, in order to accurately perform the abnormal detection operation on the commodity to be detected, the commodity completion service can be utilized to perform the information completion operation in the external inventory/price system according to the ID information, and particularly, the necessary detected field of the commodity to be detected can be pulled out according to the ID information, so that the completion result corresponding to the commodity to be detected can be generated. The necessary inspected field may include at least one of commodity description information, a commodity main map, commodity inventory, commodity price, commodity post-preferential price, and the like.

And 3, executing detection operation of the abnormal commodity on the commodity to be detected after the completion operation to obtain a preliminary detection result of the abnormal commodity.

The metadata storage service can store detection rules appointed by operators, the detection rules can comprise at least one of detection rules of commodity attribute dimension, commodity price dimension and commodity inventory dimension, the detection rules of commodity attribute dimension can comprise at least one of whether pictures/titles are matched with a delivery channel and punishment states are normal, the detection rules of commodity price dimension can comprise at least one of whether prices are matched with the delivery channel and whether specific rights can be used, and the detection rules of commodity inventory dimension can comprise at least one of whether a specific area has inventory and whether inventory values are larger than a threshold value.

In order to enable the detection operation of the abnormal commodity, the rule detection service may be used to perform the preliminary detection operation of the abnormal commodity on the commodity to be detected after the completion operation, specifically, the rule detection service may pull the detection rule from the metadata storage service, and then the rule detection service may perform the preliminary detection operation of the abnormal commodity on the commodity to be detected based on the pulled detection rule by using the avatar expression engine, so that a preliminary detection result of the abnormal commodity may be obtained.

Specifically, when the expression engine is utilized to perform preliminary detection operation on the commodity to be detected, a commodity rule expression page may be provided first, so that an operator may perform configuration operation of detection rules based on the commodity rule expression page, where the detection rules may be rules for identifying whether the commodity to be detected is in a preset attention state (for example, an on-shelf abnormal state, an off-shelf abnormal state, an abnormal state of a sales promotion activity, etc.).

And 4, the main link for abnormality detection can extract the abnormal commodity to be processed based on the preliminary detection result of the abnormal commodity.

The method comprises the steps of determining whether a commodity to be detected is in a normal state in a current detection period or not, and extracting the commodity to be detected as the commodity to be processed according to a rule detection service, wherein the rule detection service can be utilized to extract the commodity to be processed based on a preliminary detection result of the abnormal commodity, specifically, the commodity to be detected can be determined as the commodity to be processed when the preliminary detection result is used for identifying that the commodity to be detected is in the abnormal state in the current detection period, and the commodity to be detected can be determined as the commodity not to be processed according to the preliminary detection result when the commodity to be detected is in the normal state in the current detection period.

And 5, carrying out abnormal reasoning operation on the commodity to be processed to obtain the target commodity and an abnormal reasoning result.

The abnormal reasoning operation can be realized by a pre-trained abnormal commodity extraction model, and specifically, the abnormal commodity extraction model can automatically extract target concerned commodities which really need to be processed by operating personnel, wherein the target concerned commodities can be to-be-processed commodities with abnormal states continuously existing for a long time.

In some examples, the abnormal commodity extraction model may be a model implemented based on a poisson distribution function, so as to improve the instantaneity of the abnormal reasoning operation and help operators to extract target commodity concerned with the existence of an abnormal state continuously, the abnormal commodity extraction model may be used to perform the abnormal reasoning operation on the abnormal commodity to be processed, specifically, on the premise of not losing generality, it may be provided that the commodity becomes an abnormal state as a random process at a certain moment, and is irrelevant to the state at the last moment, so that the state at the next moment satisfies the poisson distribution. The abnormal commodity extraction model is used for carrying out abnormal reasoning operation on the abnormal commodity to be processed, and the abnormal reasoning operation can comprise the following steps:

step 51, calculating total anomaly times TotalAnomalies corresponding to the commodity set to be processed of the anomaly commodity and total detection period number TotalPeriods corresponding to the commodity data, and then calculating average anomaly occurrence rate by using the total anomaly times TotalAnomalies and the total detection period number TotalPeriods, specifically, average anomaly occurrence rate=total anomaly times/total detection period number.

Step 52, using the poisson distribution percentile function PoissonPPF to analyze the average anomaly occurrence rate and the preset execution level to calculate a dynamic threshold. The percentile function PoissonPPF may include, among other things, a cumulative distribution function CDF of poisson distribution and a probability mass function of poisson distribution (Probability Mass Function, PMF for short).

Step 53, for each abnormal commodity to be processed, judging whether the number of continuous abnormality times in the abnormal state sequence reaches or exceeds the dynamic threshold, if so, marking the abnormal commodity to be processed as continuous abnormality, and further determining the abnormal commodity to be processed as a target commodity of interest.

And 6, exporting the abnormal reasoning result.

The result rendering service can be utilized to conduct export operation on the abnormal reasoning results, abnormal reasoning results corresponding to the target attention commodity are obtained, and the detection result storage service can be utilized to conduct storage operation on the abnormal reasoning results. The abnormal reasoning result can be output by the detection result notification service and notified to operators, for example, the abnormal reasoning result can be sent to operators or merchant users through an instant messaging program, so that the operators or merchant users can quickly locate specific target concerned commodities based on the abnormal reasoning result, and further operate the target concerned commodities, and therefore the operation quality and effect of the target concerned commodities can be guaranteed.

Further, in addition to obtaining the abnormal reasoning result corresponding to the target commodity of interest, the result rendering service may be used to perform rendering operation on the abnormal reasoning result, specifically, a detection result report may be generated based on the large language model LLM, where the detection result report may include at least one of a curve table, a histogram, a pie chart, and the like, and the detection result report may be output, so that an operator may perform viewing operation on the detection result report. And, the index calculation service may be used to perform index calculation on the target commodity of interest, so that commodity index data corresponding to each index may be obtained, and then the commodity index data may be displayed so as to obtain commodity index data corresponding to each index corresponding to the target commodity of interest.

Further, not only can the abnormal reasoning result corresponding to the target commodity of interest be obtained, but also the commodity trend corresponding to the commodity set corresponding to the target commodity of interest can be generated through the LLM, specifically, the LLM can obtain the prompt word for generating the commodity trend based on the LLM prompt storage service through the data completion service, and then the reasoning operation can be performed on commodity data corresponding to the commodity set based on the prompt word, wherein the commodity data can comprise historical commodity data corresponding to the commodity set in a historical detection period and current commodity data corresponding to the commodity set in a current detection period, so that the commodity trend corresponding to the commodity set output by the LLM can be obtained.

In addition, after acquiring the commodity trend corresponding to the commodity set, the commodity trend corresponding to the commodity set can be analyzed and processed by utilizing the LLM, specifically, the current detection result corresponding to the commodity set in the current detection period, the statistical data of each dimension corresponding to the commodity set, the current detection reason, the historical detection result of the historical detection period corresponding to the commodity set and the historical detection reason are acquired, then the current detection result corresponding to the current detection period, the statistical data of each dimension corresponding to the commodity set, the current detection reason, the historical detection result of the historical detection period corresponding to the commodity set and the historical detection reason are input into the LLM model for analysis and processing, the operation suggestion corresponding to the commodity set, which is output by the LLM, and then the operation suggestion can be output and displayed, so that operators and merchants can perform adjustment operation of commodity operation based on the operation suggestion, and the commodity operation effect in the commodity set is improved.

The technical scheme provided by the application embodiment realizes a more advanced and intelligent abnormality detection system, can dynamically adapt to market changes, provides more accurate abnormality identification and management support, and can realize the following effects:

(1) The false alarm rate is reduced, short-term and invalid abnormal results are filtered by adopting an abnormal detection model realized by a mathematical model, so that the determined dynamic threshold value can dynamically change along with the change of commodity data and a detection period, a dynamic filtering mechanism is realized, the false alarm rate of commodity detection is further reduced, the abnormality which really needs to be concerned is separated from irrelevant information, the workload of an operation team is reduced, and the working efficiency and the decision accuracy are improved. In addition, the intelligent abnormal feedback and processing operation is realized, a full-automatic flow from detection to feedback is specifically established, operators are informed of processing in time, efficient operation of the whole link is ensured, commodity operation efficiency of the whole link is remarkably improved, higher competitiveness can be brought to an e-commerce platform, a more efficient and intelligent task management mode is provided, operation efficiency and user experience of the whole link are improved, and more accurate management support is provided for an operation team.

(2) The method breaks through dimension and scale limitations, realizes more efficient and intelligent commodity abnormality detection operation through advanced algorithm and LLM processing capability and a detection mechanism combined with rules, and can simultaneously process large-scale commodity and multidimensional data, thereby providing more comprehensive abnormality detection results, improving accuracy and flexibility of commodity abnormality detection operation, enhancing the perceptibility of the change value of the whole commodity based on the comprehensive abnormality detection results, ensuring the comprehensiveness and accuracy of the detection results, helping an operation team to make more intelligent decisions, and further improving the practicability of the scheme.

Fig. 9 is a schematic structural diagram of an abnormality detection apparatus for a commodity according to an exemplary embodiment of the present application, and referring to fig. 9, this embodiment provides an abnormality detection apparatus for a commodity, where the abnormality detection apparatus is configured to perform the abnormality detection method shown in fig. 2, and specifically the abnormality detection apparatus may include:

a first obtaining module 11, configured to obtain a commodity to be detected, where the commodity to be detected is in an abnormal state in a current detection period;

The first determining module 12 is configured to determine a number of detection periods in which the article to be detected is continuously in an abnormal state, and a dynamic threshold value corresponding to the current detection period of the article to be detected;

The first processing module 13 is configured to perform anomaly detection on the commodity to be detected based on the number of detection periods and the dynamic threshold.

The description of the embodiments shown in fig. 1 to 8 can also be performed with respect to the abnormality detection device in this embodiment, and specific reference is made to the detailed description of the embodiments, which will not be explained in detail here.

In addition, in some of the flows described in the above embodiments and the drawings, a plurality of operations appearing in a specific order are included, but it should be clearly understood that the operations may be performed out of the order in which they appear herein or performed in parallel, the sequence numbers of the operations such as 11, 12, etc. are merely used to distinguish between the various operations, and the sequence numbers themselves do not represent any order of execution. In addition, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first" and "second" herein are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, and are not limited to the "first" and the "second" being different types.

Fig. 10 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application, and as shown in fig. 10, this embodiment provides an electronic device for executing the method for detecting an abnormality of a commodity shown in fig. 2, where the electronic device may include a memory 24 and a processor 25.

Memory 24 is used to store computer programs and may be configured to store various other data to support operations on the electronic device. Examples of such data include instructions, data structures, contact data, phonebook data, messages, pictures, videos, etc. for any application or method operating on the electronic device.

The processor 25 is coupled to the memory 24 and is configured to execute the computer program in the memory 24, and is configured to obtain the commodity to be detected, wherein the commodity to be detected is in an abnormal state in the current detection period, determine the number of detection periods in which the commodity to be detected is in the abnormal state continuously and a dynamic threshold value corresponding to the commodity to be detected in the current detection period, and perform abnormal detection on the commodity to be detected based on the number of detection periods and the dynamic threshold value.

In some examples, when the processor 25 determines a dynamic threshold corresponding to a current detection period of the commodity to be detected, the processor 25 is configured to perform obtaining a detection commodity set to which the commodity to be detected belongs, determining an average anomaly occurrence rate corresponding to the detection commodity set within a preset time period, where the preset time period is greater than or equal to at least two detection periods, and determining the dynamic threshold based on the average anomaly occurrence rate and a preset confidence.

In some examples, when the processor 25 determines an average anomaly occurrence rate for the detected commodity set over the preset time period, the processor 25 is configured to obtain a total anomaly count for the detected commodity in the detected commodity set, determine a number of detection periods included in the preset time period, and determine the average anomaly occurrence rate based on the total anomaly count and the number of detection periods.

In some examples, when the processor 25 determines an average anomaly occurrence rate based on the total number of anomalies and the number of detection cycles, the processor 25 is configured to determine a ratio of the total number of anomalies to the number of detection cycles as the average anomaly occurrence rate.

In some examples, when the processor 25 determines the average anomaly occurrence rate based on the total anomaly number and the number of detection cycles, the processor 25 is configured to obtain an occurrence rate determination model, and process the total anomaly number and the number of detection cycles using the occurrence rate determination model to determine the average anomaly occurrence rate.

In some examples, when the processor 25 determines the dynamic threshold based on the average anomaly occurrence rate and the preset confidence, the processor 25 is configured to obtain a cumulative distribution function of the poisson distribution, and process the average anomaly occurrence rate and the preset confidence with the cumulative distribution function to obtain the dynamic threshold output by the cumulative distribution function.

In some examples, when the processor 25 performs the anomaly detection of the commodity to be detected based on the number of detection cycles and the dynamic threshold, the processor 25 is configured to perform determining that the commodity to be detected is an anomalous commodity if the number of detection cycles is greater than or equal to the dynamic threshold and determining that the commodity to be detected is a non-anomalous commodity if the number of detection cycles is less than the dynamic threshold.

In some examples, after determining that the article to be detected is an abnormal article, the processor 25 in this embodiment is configured to generate a detection result and an abnormal prompt for identifying the article to be detected as an abnormal article, and output the detection result and the abnormal prompt.

In some examples, when the processor 25 obtains the commodity to be detected, the processor 25 is configured to obtain a commodity candidate set, where the commodity candidate set includes commodity information corresponding to a plurality of commodity candidates, determine a detection rule for performing a preliminary detection operation on the commodity candidate set, perform an anomaly detection on the commodity candidate in the commodity candidate set by using the detection rule and the commodity information of the commodity candidate to obtain a detection result of the commodity candidate, and determine the commodity candidate corresponding to the detection result as the commodity to be detected when the detection result is used for identifying that the commodity candidate is in an anomaly state.

In some examples, when the processor 25 obtains the candidate commodity set, the processor 25 is configured to obtain commodity screening information, identify, in a preset commodity database, identity information corresponding to each of at least one candidate commodity according to the commodity screening information, and perform a completion operation of the commodity information based on the identity information corresponding to each of the at least one candidate commodity to obtain the candidate commodity set.

In some examples, after the detection of the abnormality of the commodity to be detected based on the number of detection cycles and the dynamic threshold, the processor 25 in this embodiment is configured to perform generating operation advice information corresponding to the commodity to be detected based on the abnormality detection result obtained through the abnormality detection operation, and performing association output of the abnormality detection result and the operation advice information.

In some examples, after the anomaly detection of the commodity to be detected based on the number of detection cycles and the dynamic threshold, the processor 25 in this embodiment is configured to perform the steps of obtaining a detected commodity set corresponding to the commodity to be detected, determining a current detection result of the detected commodity set in the current detection cycle and a historical detection result of the detected commodity set in the historical detection cycle, and generating an anomaly commodity trend corresponding to the detected commodity set based on the current detection result and the historical detection result.

In some examples, after generating the abnormal commodity trend corresponding to the detected commodity set, the processor 25 in this embodiment is configured to perform generating operation advice information corresponding to the detected commodity set based on the abnormal commodity trend, and outputting the abnormal commodity trend in association with the operation advice information.

Further, as shown in FIG. 10, the electronic device also includes a communication component 26, a display 27, a power supply component 28, an audio component 29, and other components. Only some of the components are schematically shown in fig. 10, which does not mean that the electronic device only comprises the components shown in fig. 10. In addition, the components within the line box in FIG. 10 are optional components, and not necessarily optional components, depending on the product form of the working node. The working node of the embodiment can be implemented as terminal equipment such as a desktop computer, a notebook computer, a smart phone or an IOT device, and also can be a server-side device such as a conventional server, a cloud server or a server array. If the working node of the embodiment is implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, etc., the working node may include the components in the line frame in fig. 10, and if the working node of the embodiment is implemented as a server device such as a conventional server, a cloud server, or a server array, etc., the working node may not include the components in the line frame in fig. 10.

The Memory may be implemented by any type or combination of volatile or non-volatile Memory devices, such as Static Random-Access Memory (SRAM), electrically erasable programmable Read-Only Memory (ELECTRICALLY ERASABLE PROGRAMMABLE READ ONLY MEMORY EEPROM), erasable programmable Read-Only Memory (Erasable Programmable Read Only Memory EPROM), programmable Read-Only Memory (Programmable Read-Only Memory, PROM), read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk.

The communication component is configured to facilitate wired or wireless communication between the device in which the communication component is located and other devices. The device where the communication component is located can access a wireless network based on a communication standard, such as a mobile communication network of 2G, 3G, 4G/LTE, 5G, etc., or a combination thereof. In one exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel.

The display includes a screen, which may include a Liquid Crystal Display (LCD) and a touch panel (TouchPanel, TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or sliding action, but also the duration and pressure associated with the touch or sliding operation.

The power supply component provides power for various components of equipment where the power supply component is located. The power components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the devices in which the power components are located.

The audio component described above may be configured to output and/or input an audio signal. For example, the audio component includes a Microphone (MIC) configured to receive external audio signals when the device in which the audio component is located is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may be further stored in a memory or transmitted via a communication component. In some embodiments, the audio assembly further comprises a speaker for outputting audio signals.

Accordingly, embodiments of the present application also provide a computer-readable storage medium storing a computer program, which when executed by a processor, causes the processor to implement the steps of the method embodiments described above. Wherein the computer-readable storage medium includes volatile or nonvolatile or a combination thereof implementations, which may be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change Memory (Phase-change Random Access Memory, PRAM for short), static Random Access Memory (SRAM), dynamic Random Access Memory (Dynamic Random Access Memory, DRAM for short), other types of Random Access Memory (RAM for short), read-only Memory (ROM), electrically erasable programmable read-only Memory (EEPROM), erasable programmable read-only Memory (EPROM), programmable read-only Memory (PROM), flash Memory or other Memory technology, compact disc read-only Memory (CD-ROM), digital versatile disc (Digital Video Disc, DVD for short) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission media

Accordingly, embodiments of the present application also provide a computer program product comprising a computer program or instructions which, when executed by a processor, cause the processor to carry out the steps of the above-described method embodiments. It should be understood that each of the above-described method flows or a combination of flows may be implemented by a computer program or instructions. In addition, these computer programs or instructions may be applied to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus, such that the processor of the general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus is implemented as a device that performs the corresponding functions in the above described method embodiments.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.

The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.

Claims

1. A method for detecting anomalies in commodities, comprising:

Acquire a commodity to be inspected, wherein the commodity to be inspected is in an abnormal state in a current inspection cycle;

Determine the number of detection cycles in which the product to be detected is in an abnormal state continuously and the dynamic threshold corresponding to the product to be detected in the current detection cycle;

Anomaly detection is performed on the commodity to be detected based on the number of detection cycles and the dynamic threshold.

2. The method according to claim 1, wherein determining the dynamic threshold corresponding to the commodity to be inspected in the current inspection cycle comprises:

Obtaining the test product set to which the product to be tested belongs;

Determining an average abnormality occurrence rate corresponding to the set of detected products within a preset time period, wherein the preset time period is greater than or equal to at least two detection cycles;

The dynamic threshold is determined based on the average abnormality occurrence rate and a preset confidence level.

3. The method according to claim 2, wherein determining the average abnormality occurrence rate corresponding to the set of detected products within a preset time period comprises:

Obtaining the total number of abnormalities corresponding to the test products in the test product set;

Determining the number of detection cycles included in the preset time period;

The average abnormality occurrence rate is determined based on the total number of abnormalities and the number of detection cycles.

4. The method according to claim 3, wherein determining the average abnormality occurrence rate based on the total number of abnormalities and the number of detection cycles comprises:

The ratio of the total number of abnormalities to the number of detection cycles is determined as the average abnormality occurrence rate.

5. The method according to claim 3, wherein determining the average abnormality occurrence rate based on the total number of abnormalities and the number of detection cycles comprises:

Obtaining incidence determination models;

The total number of abnormalities and the number of detection cycles are processed using the occurrence rate determination model to determine the average abnormality occurrence rate.

6. The method according to claim 2, wherein determining the dynamic threshold based on the average abnormality occurrence rate and a preset confidence level comprises:

Get the cumulative distribution function of the Poisson distribution;

The average abnormality occurrence rate and the preset confidence level are processed using the cumulative distribution function to obtain a dynamic threshold output by the cumulative distribution function.

7. The method according to any one of claims 1 to 6, wherein the step of performing abnormality detection on the commodity to be detected based on the number of detection cycles and the dynamic threshold comprises:

When the number of detection cycles is greater than or equal to the dynamic threshold, determining that the commodity to be detected is an abnormal commodity;

When the number of detection cycles is less than the dynamic threshold, it is determined that the commodity to be detected is a non-abnormal commodity.

8. The method according to claim 7, characterized in that after determining that the commodity to be detected is an abnormal commodity, the method further comprises:

Generate a detection result and abnormal prompt information for identifying the commodity to be detected as an abnormal commodity;

Output the detection results and abnormal prompt information.

9. The method according to any one of claims 1 to 6, wherein obtaining the commodity to be inspected comprises:

Acquire a candidate product set, wherein the candidate product set includes multiple candidate products, and the candidate products correspond to product information;

Determining a detection rule for performing a preliminary detection operation on the candidate product set;

Performing anomaly detection on the candidate products in the candidate product set using the detection rule and the product information of the candidate products to obtain a detection result of the candidate products;

In a case where the detection result is used to identify that the candidate commodity is in an abnormal state, the candidate commodity corresponding to the detection result is determined as the commodity to be detected.

10. The method according to claim 9, wherein obtaining a set of candidate products comprises:

Get product screening information;

Identifying identity information corresponding to at least one candidate product that meets the product screening information in a preset product database;

The product information is completed based on the identity identification information corresponding to each of the at least one candidate product to obtain the candidate product set.

11. The method according to any one of claims 1 to 6, characterized in that after performing abnormality detection on the commodity to be detected based on the number of detection cycles and the dynamic threshold, the method further comprises:

Based on the anomaly detection results obtained through the anomaly detection operation, generating operation suggestion information corresponding to the product to be detected;

The anomaly detection result and the operation suggestion information are correlated and output.

12. The method according to any one of claims 1 to 6, characterized in that after performing abnormality detection on the commodity to be detected based on the number of detection cycles and the dynamic threshold, the method further comprises:

Obtaining a set of test products corresponding to the product to be tested;

Determine the current test results of the test product set in the current test cycle and the historical test results in the historical test cycles;

Based on the current detection result and the historical detection results, an abnormal commodity trend corresponding to the detected commodity set is generated.

13. The method according to claim 12, characterized in that after generating the abnormal commodity trend corresponding to the detected commodity set, the method further comprises:

Based on the abnormal commodity trend, generating operation suggestion information corresponding to the detected commodity set;

The abnormal product trend is associated with the operation suggestion information and outputted.

14. A device for detecting anomalies in commodities, comprising:

A first acquisition module is used to acquire a commodity to be detected, wherein the commodity to be detected is in an abnormal state in a current detection cycle;

A first determination module is configured to determine the number of detection cycles in which the commodity to be detected is continuously in an abnormal state and a dynamic threshold corresponding to the commodity to be detected in a current detection cycle;

The first processing module is configured to perform abnormality detection on the commodity to be detected based on the number of detection cycles and the dynamic threshold.

15. An electronic device, characterized in that it comprises: a memory and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions, when executed by the processor, implement the method of any one of claims 1 to 13.

16. A computer storage medium, characterized in that it is used to store a computer program, wherein the computer program enables a computer to implement the method according to any one of claims 1 to 13 when executed.

17. A computer program product, characterized in that it comprises: a computer program, which, when executed by a processor of an electronic device, causes the processor to execute the steps of the method according to any one of claims 1 to 13.