CN110909241B

CN110909241B - Information recommendation method, user identification recommendation method, device and equipment

Info

Publication number: CN110909241B
Application number: CN201911143350.2A
Authority: CN
Inventors: 金诚; 何峰; 刘艳
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2019-11-20
Filing date: 2019-11-20
Publication date: 2023-04-07
Anticipated expiration: 2039-11-20
Also published as: CN110909241A

Abstract

The embodiment of the application discloses an information recommendation method, a user identification recommendation method, a device and equipment, and belongs to the technical field of networks. The method comprises the following steps: determining at least one second user identifier to be recommended for the first user identifier, selecting one reason generation algorithm from multiple reason generation algorithms for each second user identifier, generating a recommendation reason for the second user identifier by adopting the selected reason generation algorithm, generating user information according to the recommendation reasons of the at least one second user identifier and the at least one second user identifier, and recommending the user information to the first user identifier. By generating the recommendation reason for each user identifier to be recommended, the user can deepen the understanding of the user on the recommended user identifier, and the attraction of the user is improved, so that the association relationship between the recommended user identifier and the recommended user identifier is promoted to be established by the user, and the recommendation effect is improved.

Description

Information recommendation method, user identification recommendation method, device and equipment

Technical Field

The embodiment of the application relates to the technical field of networks, in particular to an information recommendation method, a user identification recommendation method, a device and equipment.

Background

With the development of internet technology and the increasing scale of network data, the demands of users are more and more diversified and personalized. In order to enrich the social experience of the user, the social relationship of the user is expanded, and friends are often recommended to the user.

The related art provides an information recommendation method, which includes the steps of obtaining attribute information of a first user identifier, obtaining at least one second user identifier according to the attribute information of the first user identifier based on a preset recommendation algorithm, recommending the at least one second user identifier to the first user identifier, enabling the recommended at least one second user identifier to be viewed by the subsequent first user identifier, and displaying the recommended second user identifier as shown in fig. 1. However, the above recommendation method is fixed and simple, and the recommendation effect is poor.

Disclosure of Invention

The embodiment of the application provides an information recommendation method, a user identification recommendation method, a device and equipment, which can effectively improve the attraction and recommendation effect to users. The technical scheme is as follows:

in one aspect, an information recommendation method is provided, and the method includes:

determining at least one second user identifier to be recommended for the first user identifier;

for each second user identification, selecting one reason generation algorithm from multiple reason generation algorithms, and generating a recommendation reason for the second user identification by adopting the selected reason generation algorithm;

generating user information according to the at least one second user identifier and the recommendation reason of the at least one second user identifier;

recommending the user information to the first user identification.

Optionally, the determining at least one second user identifier to be recommended for the first user identifier includes:

determining a plurality of recommendation algorithms matched with the user tags according to the user tags of the first user identifier;

and selecting one recommendation algorithm from the plurality of recommendation algorithms, and determining the at least one second user identifier for the first user identifier by adopting the selected recommendation algorithm.

Optionally, the selecting a recommendation algorithm from the multiple recommendation algorithms includes:

and selecting one recommendation algorithm according to the probability of each recommendation algorithm in the plurality of recommendation algorithms so as to match the selection times of each recommendation algorithm with the probability.

Optionally, after recommending the user information to the first user identifier, the method further includes:

and adjusting the probability of each recommendation algorithm according to the attention degree of the first user identification to the user information so as to match the probability of each recommendation algorithm with the corresponding attention degree.

determining a second user identifier set to be recommended for the first user identifier by respectively adopting each recommendation algorithm in the multiple recommendation algorithms, wherein the second user identifier set comprises at least one second user identifier;

and fusing the plurality of determined second user identification sets to obtain at least one second user identification recommended for the first user identification.

Optionally, before determining, according to a plurality of user tags of the first user identifier, a plurality of recommendation algorithms matching the plurality of user tags, the method further includes:

acquiring attribute information of the first user identifier;

and determining the plurality of user tags of the first user identification according to the attribute information of the first user identification.

Optionally, for each second subscriber identity, selecting a reason generation algorithm from a plurality of reason generation algorithms, including:

and for each second user identification, when the first user identification and the second user identification meet the triggering condition of any reason generation algorithm, selecting the any reason generation algorithm.

Optionally, the generating a recommendation reason for the second user identifier by using the selected reason generating algorithm includes:

and when the reason generation algorithm is a first reason generation algorithm, selecting a recommendation reason matched with the attribute information from a plurality of fixed recommendation reasons as the recommendation reason of the second user identification according to the attribute information of the second user identification.

and when the reason generation algorithm is a second reason generation algorithm, adding attribute information of the user identifications associated with the first user identification and the second user identification in a recommendation reason template to generate a common friend recommendation reason.

In another aspect, a user identity recommendation method is provided, where the method includes:

determining a plurality of recommendation algorithms matched with a plurality of user labels according to the plurality of user labels of the first user identification;

selecting one recommendation algorithm according to the probability of each recommendation algorithm in the plurality of recommendation algorithms so as to match the selection times of each recommendation algorithm with the probability;

determining at least one second user identifier to be recommended for the first user identifier by adopting a selected recommendation algorithm;

recommending the at least one second user identification to the first user identification.

In another aspect, an information recommendation apparatus is provided, the apparatus including:

the user identification determining module is used for determining at least one second user identification to be recommended for the first user identification;

the recommendation reason generating module is used for selecting one reason generating algorithm from multiple reason generating algorithms for each second user identifier, and generating a recommendation reason for the second user identifier by adopting the selected reason generating algorithm;

the user information generating module is used for generating user information according to the at least one second user identifier and the recommendation reason of the at least one second user identifier;

and the user information recommending module is used for recommending the user information to the first user identification.

Optionally, the subscriber identity determining module includes:

the first recommendation algorithm determining unit is used for determining multiple recommendation algorithms matched with the multiple user tags according to the multiple user tags of the first user identifier;

and the first user identifier determining unit is used for selecting one recommendation algorithm from the plurality of recommendation algorithms and determining the at least one second user identifier for the first user identifier by adopting the selected recommendation algorithm.

Optionally, the user identifier determining unit is further configured to select one recommendation algorithm according to the probability of each recommendation algorithm in the multiple recommendation algorithms, so that the selection times of each recommendation algorithm are matched with the probability.

Optionally, the apparatus further comprises:

and the first probability adjusting module is used for adjusting the probability of each recommendation algorithm according to the attention degree of the first user identification to the user information so as to enable the probability of each recommendation algorithm to be matched with the corresponding attention degree.

Optionally, the user identifier determining module includes:

the second recommendation algorithm determining unit is used for determining multiple recommendation algorithms matched with the multiple user tags according to the multiple user tags of the first user identifier;

a second user identifier determining unit, configured to determine, for the first user identifier, a second user identifier set to be recommended by using each of the multiple recommendation algorithms, where the second user identifier set includes at least one second user identifier;

and the user identifier fusion unit is used for fusing the plurality of determined second user identifier sets to obtain at least one second user identifier recommended for the first user identifier.

Optionally, the apparatus further comprises:

the attribute information acquisition module is used for acquiring the attribute information of the first user identifier;

and the user tag determining module is used for determining the plurality of user tags of the first user identifier according to the attribute information of the first user identifier.

Optionally, the recommendation reason generating module includes:

and the first generation algorithm selection unit is used for selecting any reason generation algorithm when the first user identifier and the second user identifier meet the triggering condition of the any reason generation algorithm for each second user identifier.

Optionally, the recommendation reason generating module includes:

and the second generation algorithm selection unit is used for selecting one reason generation algorithm according to the probability of each reason generation algorithm in the multiple reason generation algorithms so as to enable the selection times of each reason generation algorithm to be matched with the probability.

Optionally, the apparatus further comprises:

and the second probability adjusting module is used for adjusting the probabilities of the multiple reason generating algorithms according to the attention of the first user identifier to the user information so as to enable the probability of each reason generating algorithm to be matched with the corresponding attention.

Optionally, the recommendation reason generating module includes:

and when the reason generation algorithm is a first reason generation algorithm, the first recommendation reason selection unit is used for selecting a recommendation reason matched with the attribute information from a plurality of fixed recommendation reasons as the recommendation reason of the second user identifier according to the attribute information of the second user identifier.

Optionally, the recommendation reason generating module includes:

and the second recommendation reason selecting unit is used for adding the attribute information of the user identifications associated with the first user identification and the second user identification in a recommendation reason template to generate a common friend recommendation reason when the reason generation algorithm is the second reason generation algorithm.

Optionally, the recommendation reason generating module includes:

and the third recommendation reason selecting unit is used for generating a recommendation reason for the second user identifier based on a text generator according to the historical recommendation record of each user identifier in the category to which the first user identifier belongs when the reason generating algorithm is the third reason generating algorithm.

Optionally, the third reason recommendation selecting unit is further configured to determine, according to the historical recommendation record, a degree of attention of a recommended user identifier in the historical recommendation record to a recommendation reason, determine, according to the degree of attention of the recommended user identifier to the recommendation reason, a plurality of recommendation reasons, where the degree of attention of each recommendation reason in the plurality of recommendation reasons is greater than a preset threshold, select, according to the attribute information of the second user identifier, a recommendation reason matching the attribute information from the plurality of recommendation reasons, and determine the recommendation reason as the recommendation reason for the second user identifier.

In another aspect, an apparatus for recommending user identification is provided, the apparatus comprising:

the recommendation algorithm determining module is used for determining a plurality of recommendation algorithms matched with the user tags according to the user tags of the first user identification;

the recommendation algorithm selecting module is used for selecting one recommendation algorithm according to the probability of each recommendation algorithm in the plurality of recommendation algorithms so as to match the selection times of each recommendation algorithm with the probability;

the user identification determining module is used for determining at least one second user identification to be recommended for the first user identification by adopting a selected recommendation algorithm;

and the user identifier recommending module is used for recommending the at least one second user identifier to the first user identifier.

Optionally, the apparatus further comprises:

and the probability adjusting module is used for adjusting the probability of each recommendation algorithm according to the attention degree of the first user identifier to the at least one second user identifier so as to enable the probability of each recommendation algorithm to be matched with the corresponding attention degree.

In another aspect, a computer device is provided, which includes a processor and a memory, where at least one program code is stored, and the at least one program code is loaded and executed by the processor to implement the information recommendation method according to the above aspect; or to implement the user identification recommendation method as described in the above aspect.

In another aspect, a computer-readable storage medium is provided, in which at least one program code is stored, the at least one program code being loaded and executed by a processor to implement the information recommendation method according to the above aspect; or to implement the user identification recommendation method as described in the above aspect.

The beneficial effects that technical scheme that this application embodiment brought include at least:

the method, the device and the equipment provided by the embodiment of the application determine multiple recommendation algorithms matched with the multiple user tags according to the multiple user tags of the first user identifier, select one recommendation algorithm according to the probability of each recommendation algorithm in the multiple recommendation algorithms so that the selection times of each recommendation algorithm are matched with the probability, determine at least one second user identifier to be recommended for the first user identifier by adopting the selected recommendation algorithm, and recommend the at least one second user identifier to the first user identifier. The method for selecting the recommendation algorithm according to the probability avoids the problem that repeated recommendation effects are generated by recommending for multiple times by adopting the same recommendation algorithm, enriches the recommendation modes, improves the attraction to users and further improves the recommendation effect.

The method, the device and the equipment provided by the embodiment of the application determine at least one second user identifier to be recommended for the first user identifier, select one reason generation algorithm from multiple reason generation algorithms for each second user identifier, generate a recommendation reason for the second user identifier by adopting the selected reason generation algorithm, generate user information according to the recommendation reason of the at least one second user identifier and the at least one second user identifier, and recommend the user information to the first user identifier. By generating the recommendation reason for each user identifier to be recommended, the user can be deepened to know the recommended user identifier, the attraction to the user is improved, the association relationship between the recommended user identifier and the recommended user identifier is promoted to be established by the user, and the recommendation effect is improved.

By setting multiple recommendation algorithms and multiple reason generation algorithms, the recommendation algorithms and the reason generation algorithms are respectively selected according to the probability of each recommendation algorithm and the probability of each reason generation algorithm, and the multiple recommendation algorithms and the multiple reason generation algorithms are matched for use, so that the recommendation effect can be improved, the recommended users can obtain better social experience, the same users and the same recommendation reasons are prevented from being recommended to the users all the time, tedious user information is prevented from being recommended, and the user information has a stable recommendation effect. After the user information is recommended to the user identification, the probability of the recommendation algorithm and the probability of the reason generation algorithm can be adjusted through the attention degree of the user identification to the user information, the proportion among the recommendation algorithms and the proportion among the reason generation algorithms are adjusted, and the proportion of the recommendation algorithm and the reason generation algorithm with poor recommendation effect is reduced.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present application, and it is obvious for those skilled in the art that other drawings may be obtained according to the drawings without creative efforts.

FIG. 1 is a diagram illustrating a second user identifier of a recommendation in the related art;

FIG. 2 is a schematic illustration of an implementation environment provided by an embodiment of the present application;

fig. 3 is a flowchart of a user identifier recommendation method according to an embodiment of the present application;

fig. 4 is a flowchart of an information recommendation method according to an embodiment of the present application;

FIG. 5 is a diagram illustrating a recommended second user identifier according to an embodiment of the present disclosure;

FIG. 6 is a schematic structural diagram of a computer device according to an embodiment of the present disclosure;

fig. 7 is a schematic structural diagram of an information recommendation device according to an embodiment of the present application;

fig. 8 is a schematic structural diagram of an information recommendation device according to an embodiment of the present application;

fig. 9 is a schematic structural diagram of a user identification recommending apparatus according to an embodiment of the present application;

fig. 10 is a schematic structural diagram of a user identification recommending apparatus according to an embodiment of the present application;

fig. 11 is a schematic structural diagram of a terminal according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of a server according to an embodiment of the present application.

Detailed Description

To make the objects, technical solutions and advantages of the embodiments of the present application more clear, the embodiments of the present application will be further described in detail with reference to the accompanying drawings.

The information recommendation method provided by the embodiment of the application can be used in computer equipment.

In one possible implementation, the computer device may be a terminal, and the terminal may be a mobile phone, a computer, a tablet computer, or other types of terminals.

The terminal determines at least one second user identifier to be recommended for the first user identifier, generates a recommendation reason for each second user identifier, generates user information according to the recommendation reasons of the at least one second user identifier and the at least one second user identifier, and displays the user information based on the first user identifier, so that the user information is recommended to the user, and the user can see the recommended user information.

In another possible implementation, the computer device may include a server and a terminal.

Fig. 2 is a schematic structural diagram of an implementation environment provided in an embodiment of the present application, and as shown in fig. 2, the implementation environment includes a terminal 201 and a server 202. The terminal 201 establishes a communication connection with the server 202, and performs interaction through the established communication connection.

The terminal 201 may be a mobile phone, a computer, a tablet computer, or other types of terminals 201. The server 202 may be a server, a server cluster composed of several servers, or a cloud computing server center.

The server 202 determines at least one second user identifier to be recommended for the first user identifier, generates a recommendation reason for each second user identifier, generates user information according to the recommendation reasons of the at least one second user identifier and the at least one second user identifier, the server 202 sends the user information to the terminal 201, and the terminal 201 displays the user information so that the user can view the recommended user information.

The information recommendation method provided by the embodiment of the application can be applied to any scene of recommendation information.

For example, in a friend scenario:

the instant messaging application logs in based on a first user identifier, when a user opens a friend display interface in the instant messaging application, the server recommends user information to the first user identifier by adopting the information recommendation method provided by the embodiment of the application, the terminal displays the user information, the user information can comprise a second user identifier recommended for the first user identifier and a recommendation reason, the user can check the recommendation reason, select the second user identifier and add the second user identifier as a friend.

The embodiment of the application provides a user identifier recommendation method, which can select one recommendation algorithm from multiple recommendation algorithms, and recommend a user identifier by adopting the recommendation algorithm.

Fig. 3 is a flowchart of a user identifier recommendation method provided in an embodiment of the present application, and is applied to a computer device, as shown in fig. 3, the method includes:

301. the computer device determines a plurality of recommendation algorithms matched with the plurality of user tags according to the plurality of user tags of the first user identification.

The user identifier may be a telephone number, a user account, a user nickname, or the like, which can determine the identifier of the unique user. The user tags are used for representing categories to which the user belongs, each user tag corresponds to one dimension, and the dimensions corresponding to different user tags are different, so that each user identifier can have one user tag or can also have a plurality of user tags at the same time. For example, the user tag may include a social tag, a paid tag, a preference tag, and the like, for example, if the number of other user tags associated with the user tag is large and interactive chat is performed with the associated other user tags frequently, it may be determined that the user tag has the social tag; the payment behavior records of the user identifier are multiple, and the user identifier can be determined to have a payment label; the user identification often browses news, material, etc. about the movie, it may be determined that the user identification has a movie preference tag.

The recommendation algorithm is an algorithm for recommending other user identifiers for the user identifier according to the attribute information of the user identifier, such as a graph mining algorithm, an affinity propagation algorithm, a PageRank algorithm, a Network Embedding algorithm, a random recommendation algorithm, an AI (Artificial Intelligence) recommendation algorithm based on a user avatar picture, a knowledge graph, a social Network and other fused graph convolution algorithms, an Influence Maximization algorithm and the like.

Different recommendation algorithms have different use conditions, for example, the AI recommendation algorithm based on the user avatar picture is applicable to the case that avatar pictures of multiple user identifiers can be acquired, and the recommendation algorithm cannot be used for the user identifier that cannot acquire the user avatar picture.

Therefore, a matching user label is set for each recommendation algorithm with respect to the use condition of each recommendation algorithm. When recommending a user identifier for the first user identifier, a recommendation algorithm matching the user identifier may be determined according to the user identifier of the first user identifier.

It should be noted that in the embodiment of the present application, a plurality of recommendation algorithms are determined based on determining a plurality of user tags of the first user identifier, and before this step 301, the computer device obtains the user tags of the first user identifier.

For the process of obtaining the user tag of the first user identifier, in a possible implementation manner, the computer device obtains attribute information of the first user identifier, and determines multiple user tags of the first user identifier according to the attribute information of the first user identifier.

The attribute information may include the activity level of the user, the geographical location, whether the user is a paying user, hobbies, and the like. When the attribute information of the first user identifier is obtained, the behavior record of the first user identifier can be analyzed and processed by adopting a user activity model, a user loss model, a payment portrait model, a user game map model and the like, so that the attribute information of the first user identifier is obtained, and a corresponding user label is obtained according to the attribute information.

Since the attribute information may include attribute information of multiple dimensions, such as geographic location and hobbies belonging to different dimensions, multiple user tags may be determined based on the attribute information of the first user identifying the multiple dimensions.

Optionally, according to the attribute information of the first user identifier, a classification model is used to determine a user tag to which the first user identifier belongs. The classification model comprises a plurality of preset user labels and attribute information corresponding to each user label, and the attribute information of the first user identification is input into the classification model, so that the user labels corresponding to the attribute information can be obtained.

302. The computer device selects one recommendation algorithm according to the probability of each recommendation algorithm in the plurality of recommendation algorithms so that the selection times of each recommendation algorithm are matched with the probability.

For the same user identifier, if the same recommendation algorithm is adopted every time, the user identifiers to be recommended obtained every time are the same, and the user identifiers recommended every time are the same when the recommendation is carried out on the same user identifier for multiple times, so that the user loses interest in the recommended user identifiers, and the recommendation effect is influenced.

Therefore, a plurality of recommendation algorithms are set, and a corresponding probability is set for each recommendation algorithm, wherein the probability of each recommendation algorithm represents the possibility of selecting each recommendation algorithm when the recommendation algorithm is selected each time, and the probability of the recommendation algorithm is higher, and the probability of selecting the recommendation algorithm is higher. Therefore, when recommendation is performed each time, the recommendation algorithm is selected according to the probabilities of the plurality of recommendation algorithms, so that the selection times of each recommendation algorithm are matched with the probabilities.

The selection times and the probability matching of the recommendation algorithm mean that the ratio of the selection times of the recommendation algorithm to the total times of selecting the recommendation algorithm is close to the probability of the selected recommendation algorithm.

For example, when the ratio of the number of selection times to the total number of times of a certain recommendation algorithm is greater than the probability of the recommendation algorithm, it indicates that the recommendation algorithm has been selected many times, and no selection is required this time. And when the ratio of the selection times to the total times of a certain recommended algorithm is not more than the probability of the recommended algorithm, the recommended algorithm is less selected, and the selection can be performed at this time.

For example, the user tag of the first user identification applies to four recommendation algorithms: the method comprises a recommendation algorithm A, a recommendation algorithm B, a recommendation algorithm C and a recommendation algorithm D, wherein the probability of the recommendation algorithm A is 0.2, the probability of the recommendation algorithm B is 0.3, the probability of the recommendation algorithm C is 0.1, and the probability of the recommendation algorithm D is 0.4, so that in the process of recommending the user identifier for the first user identifier 10 times, the number of selecting the recommendation algorithm A is 2, the number of selecting the recommendation algorithm B is 3, the number of selecting the recommendation algorithm C is 1, and the number of selecting the recommendation algorithm D is 4.

303. And the computer equipment determines at least one second user identifier to be recommended for the first user identifier by adopting the selected recommendation algorithm.

The recommendation algorithm is to recommend at least one second user identifier for the first user identifier after determining the relevance between different user identifiers according to the attribute information of the first user identifier, so that the at least one second user identifier can be regarded as a user identifier in which the first user identifier is interested.

304. The computer device recommends at least one second user identification to the first user identification.

The computer device recommends the at least one second user identification to the first user identification so that the first user identification can view the recommended at least one second user identification.

In a possible implementation manner, the computer device is a terminal, and the terminal logs in based on the first user identifier, and then the terminal acquires a plurality of user identifiers of the login server, and after determining at least one second user identifier to be recommended for the first user identifier by using a selected recommendation algorithm, displays the at least one second user identifier, thereby recommending the at least one second user identifier to the user.

In another possible implementation manner, the computer device is a server, and the terminal logs in the server based on the first user identifier, the server obtains a plurality of user identifiers that log in the server, determines at least one second user identifier to be recommended for the first user identifier by using a selected recommendation algorithm, and then sends the at least one second user identifier to the terminal, and the terminal displays the at least one second user identifier, so that the at least one second user identifier is recommended to the user.

When recommending the second user identifier to the first user identifier, only the second user identifier itself may be recommended, or the second user identifier and the head portrait and the nickname of the second user identifier may also be recommended.

It should be noted that, in one possible implementation, after step 304, the method further includes: and the computer equipment adjusts the probability of each recommendation algorithm according to the attention degree of the first user identification to at least one second user identification so as to enable the probability of each recommendation algorithm to be matched with the corresponding attention degree.

The attention degree is used for representing the attention degree of the first user identifier to the recommended user identifier, the higher the attention degree is, the more interesting the first user identifier is to the recommended result is, and the lower the attention degree is, the less interesting the first user identifier is to the recommended result is. For example, the attention degree is whether the first user identifier and the recommended at least one second user identifier establish an association relationship, if the user requests to establish the association relationship between the first user identifier and the at least one second user identifier according to the recommended at least one second user identifier, it indicates that the attention degree of the user to the recommended at least one second user identifier is high, and if the user does not establish an association relationship with any second user identifier, it indicates that the attention degree of the user to the recommended at least one second user identifier is low.

Therefore, the probability of each recommendation algorithm is adjusted according to the attention of the first user identifier to the recommended at least one second user identifier, when the attention of the first user identifier to the recommended at least one second user identifier is low, the probability of the recommendation algorithm corresponding to the at least one second user identifier is adjusted to be low, the probabilities of other recommendation algorithms are adjusted to be high, when the attention of the first user identifier to the recommended at least one second user identifier is high, the probability of the recommendation algorithm corresponding to the at least one second user identifier is adjusted to be high, and the probabilities of the other recommendation algorithms are adjusted to be low.

In addition, for the probability of each recommendation algorithm, in one possible implementation manner, for the multiple recommendation algorithms acquired for the first time, the probability of each recommendation algorithm is the same, and the sum of the probabilities of the multiple recommendation algorithms is a fixed value, for example, the sum of the probabilities of the multiple recommendation algorithms is 1. After any recommendation algorithm is adopted to recommend to the first user identifier, the probability of each recommendation algorithm is adjusted according to the attention degree of the first user identifier to the recommended user identifier, and the sum of the probabilities of the multiple recommendation algorithms is still the fixed value after adjustment.

According to the user identifier recommending method provided by the embodiment of the application, multiple recommending algorithms matched with the multiple user tags are determined according to the multiple user tags of the first user identifier, one recommending algorithm is selected according to the probability of each recommending algorithm in the multiple recommending algorithms, so that the selecting times of each recommending algorithm are matched with the probability, at least one second user identifier to be recommended is determined for the first user identifier by adopting the selected recommending algorithm, and the at least one second user identifier is recommended to the first user identifier. The method for selecting the recommendation algorithm according to the probability avoids the problem that repeated recommendation effects are generated by recommending for multiple times by adopting the same recommendation algorithm, enriches the recommendation modes, improves the attraction to users and further improves the recommendation effect.

It should be noted that, in the embodiment of the present application, an example of selecting one recommendation algorithm from a plurality of recommendation algorithms is taken as an example, and in another implementation, steps 302 and 303 may be replaced with the following steps: according to the multiple user tags of the first user identification, multiple recommendation algorithms matched with the multiple user tags are determined, each recommendation algorithm in the multiple recommendation algorithms is respectively adopted to determine a second user identification set to be recommended for the first user identification, and the second user identification set comprises at least one second user identification.

Because different recommendation algorithms cover different user groups, the user identifications to be recommended obtained by the different recommendation algorithms are different, and in order to enrich the user identifications to be recommended, the second user identification sets obtained by the multiple recommendation algorithms are fused, so that more user identifications can be obtained, and the recommended information amount is enriched.

Fig. 4 is a flowchart of an information recommendation method provided in an embodiment of the present application, which is applied to a computer device, and as shown in fig. 4, the method includes:

401. the computer device obtains attribute information of the first user identifier.

The step 401 is similar to the manner of obtaining the attribute information of the first ue in the step 301, and is not described herein again.

402. The computer device determines a plurality of user tags of the first user identifier according to the attribute information of the first user identifier.

This step 402 is similar to the manner of determining the multiple user tags of the first user identifier according to the attribute information of the first user identifier in step 301 in the foregoing embodiment, and is not described herein again.

403. The computer device determines a plurality of recommendation algorithms matched with the plurality of user tags according to the plurality of user tags of the first user identification.

This step 403 is similar to the step 301, and will not be described herein again.

404. The computer equipment selects one recommendation algorithm from the multiple recommendation algorithms, and determines at least one second user identifier to be recommended for the first user identifier by adopting the selected recommendation algorithm.

Step 404 is similar to step 303, and will not be described herein again.

When the computer device selects one recommendation algorithm from the multiple recommendation algorithms, the computer device may select the recommendation algorithm in a similar manner in step 302, or may select the recommendation algorithm randomly, or may select the recommendation algorithm in another manner.

It should be noted that, in the embodiment of the present application, the above steps 401 to 404 are taken as an example, and in another embodiment, the computer device does not need to perform the steps 401 to 404, and may also determine at least one second user identifier to be recommended for the first user identifier in other manners.

405. For each second subscriber identity, the computer device selects a cause generation algorithm from a plurality of cause generation algorithms.

In order to enhance the recommendation effect and improve the attraction of the at least one second user identifier to the first user identifier, for the at least one second user identifier to be recommended, a recommendation reason may be generated for each second user identifier.

The reason generation algorithm is an algorithm for generating a recommendation reason for the user identifier, and may be a fixed reason generation algorithm, a filling reason generation algorithm, a history reason generation algorithm, or the like. The reason for recommendation is used to prompt the user why the second subscriber identity is recommended, i.e. to prompt the user how the second subscriber identity is selected. For example, the reason for recommendation of the second user identifier is "active player", which means that the second user identifier is determined according to the activity level, and the second user identifier is the more active user identifier.

In one possible implementation, the manner of selecting one reason generating algorithm from a plurality of reason generating algorithms may include the following two:

the first mode is as follows: and for each second user identification, when the first user identification and the second user identification meet the triggering condition of any reason generation algorithm, selecting any reason generation algorithm.

Among the plurality of reason generation algorithms, different reason generation algorithms have different trigger conditions. The trigger condition of the reason generation algorithm is a precondition for adopting the reason generation algorithm, and the reason generation algorithm can be adopted only when the trigger condition is satisfied. For example, if the triggering condition of the common friend reason generation algorithm is that the number of common friends is greater than or equal to 1, the information filling reason generation algorithm is satisfied when the number of common friends of the first user identifier and the second user identifier is 2.

The second mode is as follows: one reason generation algorithm is selected according to the probability of each reason generation algorithm in the plurality of reason generation algorithms, so that the selection times of each reason generation algorithm are matched with the probability. The method is similar to the scheme of selecting one recommendation algorithm from the multiple recommendation algorithms in step 302 of the foregoing embodiment, and is not described herein again.

406. And the computer equipment adopts the selected reason generation algorithm to generate a recommendation reason for the second user identification.

Optionally, this step 406 may include the following three ways:

the first mode is as follows: and when the selected reason generation algorithm is the first reason generation algorithm, selecting a recommendation reason matched with the attribute information from the fixed recommendation reasons as the recommendation reason of the second user identifier according to the attribute information of the second user identifier.

The first reason generation algorithm is a fixed recommendation reason generation algorithm, and can generate a fixed recommendation reason. A plurality of fixed reasons for recommendation may be set in advance, each fixed reason for recommendation being provided with matching attribute information. For example, the fixed recommendation reasons are "active player", "related person", "level approaching", "high-end racing player", "same-segment friend making", and the like, wherein the user attribute information matched by the "active player" is a user with high activity, that is, the user is frequently logged in.

The second mode is as follows: and when the selected reason generation algorithm is the second reason generation algorithm, adding the attribute information of the user identification associated with both the first user identification and the second user identification in the recommendation reason template to generate the common friend recommendation reason.

The second reason generation algorithm is a common friend reason generation algorithm, a fixed recommendation reason template can be set, and the attribute information of the user identification is added to the recommendation reason template according to the user identification associated with the first user identification and the second user identification to generate the common friend recommendation reason.

The first user identifier and the second user identifier are both associated with user identifiers, and a common friend is identified for the first user identifier and the second user identifier, the attribute information of the user identifiers may include an avatar, a nickname, and the like of the user identifiers, and the generated common friend recommendation reason may include the number of the common friends of the first user identifier and the second user identifier, an avatar of each common friend, a nickname of each common friend, and the like.

For example, the number of common friends of the first user identifier and the second user identifier is 3, and the common friends are friend 1, friend 2, and friend 3, respectively, and the reason for generating the common friend recommendation is that "friend 1, friend 2, and friend 3 are common friends of your.

For example, the recommendation reason template is "| #" the website of the avatar picture of the user id 1| # "the website of the avatar picture of the user id 2| # is his/her friend", and the recommendation reason generated is "the avatar of the user id1 and the avatar of the user id 2 are his/her friend".

Or the recommendation reason template is the number of the user identifications associated with the first user identification and the second user identification, and the head portrait and the nickname of each user identification, namely the number of the user identifications associated with the first user identification and the second user identification, and the nickname of each user identification, wherein the number of the user identifications is equal to the number of the website address of the head portrait picture of each user identification, and the nickname of each user identification. And in order to avoid the influence of the messy codes in the nickname, the nickname can be encoded by adopting Base64 (an encoding mode), and the generated recommendation reasons are that the head portrait of the user identifier 1 and the nickname 1, the head portrait of the user identifier 2 and the nickname 2, and the head portrait of the user identifier 3 and the nickname 3 are common friends.

The third mode is as follows: and when the reason generation algorithm is a third reason generation algorithm, generating a recommendation reason for the second user identifier based on a text generation model according to the historical recommendation record of each user identifier in the category to which the first user identifier belongs.

The historical recommendation record comprises a plurality of recommended user identifications and corresponding recommendation reasons. And the third reason generation algorithm is a historical reason generation algorithm and is used for selecting at least one recommendation reason from the historical recommendation records as the recommendation reason of the second user identifier.

The Text generation model may be GAN (generic adaptive Networks, generative antagonistic network), textGAN (Text Generative antagonistic network), leak generated antagonistic network, or the like. The text generation model can be obtained by training the user identification and the corresponding recommendation reason in the historical recommendation record, and can also be obtained by training the set recommendation reason. For the training process, a plurality of user identifications in the history record, recommendation reasons corresponding to each user identification and the recommended user attention are obtained, and the text generation model is trained, so that the recommendation reasons with high attention generated for the user identifications are obtained.

Each user label corresponds to a dimension, and the user label represents a category to which the user belongs, so that each user label in the category to which the first user label belongs refers to other user labels having the same user label as the first user label, and the history recommendation records of the other user labels include user information received by the other user labels. Therefore, the third reason generation algorithm is adopted, and the recommendation reason with high attention is used as the recommendation reason of the second mark according to the user marks recommended in the historical recommendation record, the recommendation reason of each recommended user mark and the attention degree of the recommended user mark to the recommended user information.

Since the attention degrees of different user identifiers belonging to the same category to the same recommendation reason are similar, the attention degree of the user identifier belonging to the same category to the different recommendation reason can be used as the attention degree of other user identifiers belonging to the same category to the different recommendation reason. In this process, the recommended user identifier corresponding to the historical recommendation record is the same as the category to which the first user identifier belongs, so that a recommendation reason can be generated for the second user identifier according to the attention degree of the recommended user identifier to the recommendation reason in the historical recommendation record and the recommendation reason, and the recommendation reason can be regarded as the recommendation reason concerned by the user of the first user identifier.

Optionally, the specific process of generating the recommendation reason according to the historical recommendation record may include: according to the historical recommendation record, the attention degree of the recommended user identification in the historical recommendation record to the recommendation reason is determined, a plurality of recommendation reasons are determined according to the attention degree of the recommended user identification to the recommendation reason, the attention degree of each recommendation reason in the plurality of recommendation reasons is larger than a preset threshold value, according to the attribute information of the second user identification, the recommendation reason matched with the attribute information is selected from the plurality of recommendation reasons, and the recommendation reason is determined as the recommendation reason of the second user identification.

The attention degree may be a click rate of the recommended user identifier, and may be determined by counting whether an association relationship is established between the recommended user identifier and the recommended user identifier, for example, if the recommended user identifier establishes a friend relationship with the plurality of recommended user identifiers through the plurality of recommended user identifiers, the attention degree of the recommended user to a recommendation reason of the recommended user identifier establishing the friend relationship is high. In order to ensure that the second user identifier has a higher attention to the generated recommendation reason, so as to enhance the recommendation effect, a plurality of recommendation reasons with the attention greater than a preset threshold value are selected from the historical recommendation records, for example, the preset threshold value is 80%.

407. And the computer equipment generates user information according to the recommendation reasons of the at least one second user identifier and the at least one second user identifier.

The user information may be at least one second user identifier presented in a list form and a recommendation reason corresponding to each second user identifier.

In a possible implementation manner, the list includes at least one second user identifier and recommendation reasons of the at least one second user identifier, and each second user identifier and the corresponding recommendation reason are sorted according to the sequence that the attention degree of the recommendation user identifier to the recommendation reasons is from high to low. For example, the second user identifier includes a user identifier 1, a user identifier 2, and a user identifier 3, the recommendation reason corresponding to the user identifier 1 is recommendation reason a, the recommendation reason corresponding to the user identifier 2 is recommendation reason B, and the recommendation reason corresponding to the user identifier 3 is recommendation reason C, and the attention degree of the recommended user identifier for each recommendation reason is: when the degree of attention of the recommendation reason a is 80%, the degree of attention of the recommendation reason B is 70%, and the degree of attention of the recommendation reason a is 60%, the user id in the generated user information is user id1, user id 2, and user id 3 in the order of arrangement.

408. The computer device recommends user information to the first user identification.

The computer equipment recommends the user information to the first user identification, so that the first user identification can view the recommended user information, and a friend relationship can be established with the second user identification according to the recommended user information. For example, after the first user identifier receives the recommended user information, the displayed user information is as shown in fig. 5, and a plurality of second user identifiers and a recommendation reason corresponding to each second user identifier are displayed in the interface.

In a possible implementation manner, the computer device is a terminal, and the terminal logs in based on the first user identifier, the terminal acquires a plurality of user identifiers of the login server, determines at least one second user identifier to be recommended and a recommendation reason of each second user identifier for the first user identifier by using a selected recommendation algorithm and a reason generation algorithm, generates user information, and displays the at least one second user identifier and the recommendation reason of each second user identifier, so that the at least one second user identifier is recommended to the user.

In another possible implementation manner, the computer device is a server, and the terminal logs in the server based on the first user identifier, the server obtains multiple user identifiers that log in the server, determines, for the first user identifier, at least one second user identifier to be recommended and a recommendation reason for each second user identifier by using a selected recommendation algorithm and a reason generation algorithm, sends the user information to the terminal after the user information is generated, and displays, by the terminal, the at least one second user identifier and the recommendation reason for each second user identifier, thereby recommending the at least one second user identifier to the user.

In addition, the user information generated by the computer device may include at least one second user identifier to be recommended and a recommendation reason for each second user identifier, and may also include at least one second user identifier to be recommended, a reason ID (Identity Document) corresponding to each second user identifier, and auxiliary information for each reason ID. For example, when the computer device is a terminal, the user information generated by the terminal includes at least one second user identifier to be recommended and a recommendation reason for each second user identifier; when the computer equipment is a server, the user information generated by the server comprises at least one second user identifier to be recommended, a reason ID corresponding to each second user identifier and auxiliary information of each reason ID, the server recommends the user information to the terminal, and the terminal carries out analysis processing according to the reason ID and the auxiliary information of each reason ID and generates a corresponding recommendation reason for each second user identifier.

If a corresponding reason ID can be set for each reason of recommendation, and different reasons of recommendation can have different reason IDs, one reason of recommendation can be determined according to any reason ID.

For example, if the recommendation reason may have different reason IDs, the template of the recommendation reason is "| user ID 1| reason ID 1|", the template includes user ID1 and reason ID1 corresponding to reason ID1, and the reason corresponding to reason ID1 is "active player", the terminal analyzes "| user ID 1| reason ID 1|" carried in the user information, and the recommendation reason generated for user ID1 is "active player". Alternatively, when each recommendation reason may have the same reason ID, the template of the recommendation reason is "| user ID 1| reason ID 7| find best racing team |", | user ID 2| reason ID 7| high-end racing player | or the like, the terminal analyzes the template of the recommendation reason carried in the user information, generates a recommendation reason "find best racing team" for user ID1, and generates a recommendation reason "high-end racing player" for user ID 2.

It should be noted that, in a possible implementation manner, after step 407, the following two manners are further included:

the first mode is as follows: and adjusting the probability of each recommendation algorithm according to the attention degree of the first user identification to the user information so as to match the probability of each recommendation algorithm with the corresponding attention degree.

The second mode is as follows: and adjusting the probabilities of the multiple reason generation algorithms according to the attention degree of the first user identifier to the user information so as to enable the probability of each reason generation algorithm to be matched with the corresponding attention degree.

The two methods are similar to the process of adjusting the probability of each recommendation algorithm according to the attention of the first user identifier to the at least one second user identifier in the above embodiment, and are not described herein again.

The information recommendation method provided by the embodiment of the application determines at least one second user identifier to be recommended for a first user identifier, selects one reason generation algorithm from multiple reason generation algorithms for each second user identifier, generates a recommendation reason for the second user identifier by using the selected reason generation algorithm, generates user information according to the recommendation reasons of the at least one second user identifier and the at least one second user identifier, and recommends the user information to the first user identifier. By generating the recommendation reason for each user identifier to be recommended, the user can be deepened to know the recommended user identifier, the attraction to the user is improved, the association relationship between the recommended user identifier and the recommended user identifier is promoted to be established by the user, and the recommendation effect is improved.

It should be noted that the computer device may include a plurality of modules, as shown in fig. 6, the plurality of modules include: the system comprises an attribute information module, a recommendation algorithm module, a reason generation algorithm module, an integration module, a historical effect data module and a front-end module.

The attribute information module is configured to obtain attribute information of the user identifier, and may include basic attribute information of the user identifier, social attribute information, and game related attribute information, where the basic attribute information may be an age, a location, an academic calendar, an occupation, and the like corresponding to the user identifier, the social attribute information may include other user identifiers associated with the user identifier, and the game related attribute may be an activity degree, a game type preference, and the like.

The recommendation algorithm module is used for recommending other user identification algorithms for the user identification according to the attribute information of the user identification, such as a graph mining algorithm, an affinity propagation algorithm, a webpage sorting algorithm, a network embedding algorithm, a random recommendation algorithm, an artificial intelligence recommendation algorithm based on a user avatar picture, a knowledge graph, a social network and other fused graph volume algorithms and an influence maximization algorithm.

The reason generation algorithm module is used for generating an algorithm of a recommendation reason for the user identifier to be recommended according to the attribute information of the user identifier, and the reason generation algorithm can be a fixed reason generation algorithm, a social information reason generation algorithm, a game self-defined reason algorithm, a history reason generation algorithm and the like.

The history recommendation recording module is used for acquiring the attention degree to the recommendation reason in the history recommendation record and the attention degree to the user identifier recommended in the history recommendation record aiming at the history recommendation record of any user identifier, and can send the attention degree to the recommendation reason in the history recommendation record to the reason generation algorithm module, so that the reason generation algorithm module can generate the recommendation reason for the user identifier according to the attention degree of the recommendation reason, and also can send the attention degree to the user identifier recommended in the history recommendation record to the integration module, and the integration module adjusts the proportion of the recommendation algorithm and the reason generation algorithm according to the attention degree to the recommendation reason in the history recommendation record and the attention degree to the user identifier recommended in the history recommendation record.

The integration module is used for selecting a recommendation algorithm from the algorithm module, determining the user identification to be recommended, selecting a reason generation algorithm from the reason generation algorithm module, generating a recommendation reason for each user identification to be recommended, generating user information according to the user identification to be recommended and the corresponding recommendation reason, pushing the generated user information to the front-end module, and adjusting the probability of the recommendation algorithm and the probability of the reason generation algorithm.

The front-end module is used for displaying the received user information. In addition, the user information sent by the integration module comprises a user identifier to be recommended, a reason ID and an auxiliary field, the front-end module analyzes according to the reason ID and the auxiliary field to obtain a complete recommendation reason, and the user identifier to be recommended and the corresponding recommendation reason are displayed. The user can perform subsequent operation processing based on the front-end module, such as establishing a friend relationship with the recommended user identifier.

It should be noted that, in the embodiment of the present application, a recommendation algorithm and a reason generation algorithm are respectively selected as an example, and in another embodiment, a recommendation algorithm is first selected, according to the selected recommendation algorithm, at least one second user identifier to be recommended is determined, and an association relationship between the first user identifier and the at least one second user identifier can be obtained, and a recommendation reason of each second user identifier is determined according to the association relationship, and it is not necessary to select a reason generation algorithm to generate a corresponding recommendation reason for each second user identifier.

Fig. 7 is a schematic structural diagram of an information recommendation apparatus according to an embodiment of the present application, and as shown in fig. 7, the apparatus includes:

a user identifier determining module 701, configured to determine, for the first user identifier, at least one second user identifier to be recommended;

a reason for recommendation generating module 702, configured to select, for each second user identifier, one reason generating algorithm from multiple reason generating algorithms, and generate a reason for recommendation for the second user identifier by using the selected reason generating algorithm;

the user information generating module 703 is configured to generate user information according to at least one second user identifier and a recommendation reason for the at least one second user identifier;

and a user information recommending module 704, configured to recommend the user information to the first user identifier.

The information recommendation device provided by the embodiment of the application determines at least one second user identifier to be recommended for a first user identifier, selects one reason generation algorithm from multiple reason generation algorithms for each second user identifier, generates a recommendation reason for the second user identifier by using the selected reason generation algorithm, generates user information according to the recommendation reasons of the at least one second user identifier and the at least one second user identifier, and recommends the user information to the first user identifier. By generating the recommendation reason for each user identifier to be recommended, the user can be deepened to know the recommended user identifier, the attraction to the user is improved, the association relationship between the recommended user identifier and the recommended user identifier is promoted to be established by the user, and the recommendation effect is improved.

Optionally, as shown in fig. 8, the user identity determining module 701 includes:

a first recommendation algorithm determining unit 7101, configured to determine, according to the multiple user tags of the first user identifier, multiple recommendation algorithms matched with the multiple user tags;

the first subscriber identity determining unit 7102 is configured to select one recommendation algorithm from the multiple recommendation algorithms, and determine at least one second subscriber identity for the first subscriber identity by using the selected recommendation algorithm.

Optionally, as shown in fig. 8, the first user identifier determining unit 7102 is further configured to select one recommendation algorithm according to a probability of each recommendation algorithm in the plurality of recommendation algorithms, so that the selection times of each recommendation algorithm matches the probability.

Optionally, as shown in fig. 8, the apparatus further comprises:

the first probability adjusting module 705 is configured to adjust the probability of each recommendation algorithm according to the attention of the first user identifier to the user information, so that the probability of each recommendation algorithm matches with the corresponding attention.

a second recommendation algorithm determining unit 7103, configured to determine, according to the multiple user tags of the first user identifier, multiple recommendation algorithms matched with the multiple user tags;

a second user identifier determining unit 7104, configured to determine, for the first user identifier, a second user identifier set to be recommended by using each of the multiple recommendation algorithms, where the second user identifier set includes at least one second user identifier;

and a user identifier fusing unit 7105, configured to fuse the determined multiple second user identifier sets to obtain at least one second user identifier recommended for the first user identifier.

Optionally, as shown in fig. 8, the apparatus further comprises:

an attribute information obtaining module 706, configured to obtain attribute information of the first user identifier;

a user tag determining module 707, configured to determine, according to the attribute information of the first user identifier, a plurality of user tags of the first user identifier.

Optionally, as shown in fig. 8, the recommendation reason generating module 702 includes:

the first generation algorithm selecting unit 7201 is configured to select, for each second subscriber identity, any reason generation algorithm when the first subscriber identity and the second subscriber identity satisfy a trigger condition of any reason generation algorithm.

the second generation algorithm selecting unit 7202 selects one of the reason generation algorithms according to the probability of each of the plurality of reason generation algorithms so that the number of times of selection of each of the reason generation algorithms matches the probability.

Optionally, as shown in fig. 8, the apparatus further comprises:

the second probability adjusting module 708 is configured to adjust the probabilities of the multiple reason generating algorithms according to the attention of the first user identifier to the user information, so that the probability of each reason generating algorithm matches with the corresponding attention.

the first reason for recommendation selecting unit 7203 is configured to, when the reason generation algorithm is the first reason generation algorithm, select, as the reason for recommendation of the second user identifier, a reason for recommendation that matches the attribute information from the plurality of fixed reasons for recommendation according to the attribute information of the second user identifier.

the second recommendation reason selecting unit 7204 is configured to, when the reason generation algorithm is the second reason generation algorithm, add attribute information of the user identifiers associated with both the first user identifier and the second user identifier to the recommendation reason template, and generate a common friend recommendation reason.

the third reason for recommendation selecting unit 7205, configured to, when the reason generating algorithm is the third reason generating algorithm, generate a reason for recommendation for the second user identifier based on the text generator according to the historical recommendation record of each user identifier in the category to which the first user identifier belongs.

Optionally, as shown in fig. 8, the third reason for recommendation selecting unit 7205 is further configured to determine, according to the historical recommendation record, a degree of attention of the recommended user identifier in the historical recommendation record to the recommendation reason, determine, according to the degree of attention of the recommended user identifier to the recommendation reason, a plurality of recommendation reasons, where the degree of attention of each of the plurality of recommendation reasons is greater than a preset threshold, select, according to the attribute information of the second user identifier, a recommendation reason matching the attribute information from the plurality of recommendation reasons, and determine the recommendation reason as the second user identifier.

Fig. 9 is a schematic structural diagram of a user identification recommendation apparatus according to an embodiment of the present application, and as shown in fig. 9, the apparatus includes:

a recommendation algorithm determining module 901, configured to determine, according to the multiple user tags of the first user identifier, multiple recommendation algorithms matched with the multiple user tags;

a recommendation algorithm selecting module 902, configured to select one recommendation algorithm according to a probability of each recommendation algorithm in the multiple recommendation algorithms, so that the selection times of each recommendation algorithm are matched with the probability;

a user identifier determining module 903, configured to determine, for the first user identifier, at least one second user identifier to be recommended by using the selected recommendation algorithm;

and a user identifier recommending module 904, configured to recommend at least one second user identifier to the first user identifier.

The user identifier recommending device provided by the embodiment of the application determines multiple recommending algorithms matched with the multiple user tags according to the multiple user tags of the first user identifier, selects one recommending algorithm according to the probability of each recommending algorithm in the multiple recommending algorithms so that the selection times of each recommending algorithm are matched with the probability, determines at least one second user identifier to be recommended for the first user identifier by adopting the selected recommending algorithm, and recommends the at least one second user identifier to the first user identifier. The method for selecting the recommendation algorithm according to the probability avoids the problem that repeated recommendation effects are generated by recommending for multiple times by adopting the same recommendation algorithm, enriches the recommendation modes, improves the attraction to users and further improves the recommendation effect.

Optionally, as shown in fig. 10, the apparatus further comprises:

the probability adjusting module 905 is configured to adjust the probability of each recommendation algorithm according to the attention of the first user identifier to at least one second user identifier, so that the probability of each recommendation algorithm matches with the corresponding attention.

Fig. 11 is a schematic structural diagram of a terminal according to an embodiment of the present application, which can implement operations executed by a first terminal, a second terminal, and a third terminal in the foregoing embodiments. The terminal 1100 may be a portable mobile terminal such as: the mobile terminal comprises a smart phone, a tablet computer, an MP3 player (Moving Picture Experts Group Audio Layer III, moving Picture Experts compress standard Audio Layer 3), an MP4 player (Moving Picture Experts Group Audio Layer IV, moving Picture Experts compress standard Audio Layer 4), a notebook computer, a desktop computer, a head-mounted device, a smart television, a smart sound box, a smart remote controller, a smart microphone, or any other smart terminal. Terminal 1100 may also be referred to by other names such as user equipment, portable terminal, laptop terminal, desktop terminal, and so forth.

In general, terminal 1100 includes: a processor 1101 and a memory 1102.

Processor 1101 may include one or more processing cores, such as 4-core processors, 8-core processors, etc. Memory 1102 may include one or more computer-readable storage media, which may be non-transitory, for storing at least one instruction for processor 1101 to implement the information recommendation method or the user identification recommendation method provided by method embodiments herein.

In some embodiments, the terminal 1100 may further optionally include: a peripheral interface 1103 and at least one peripheral. The processor 1101, memory 1102 and peripheral interface 1103 may be connected by a bus or signal lines. Various peripheral devices may be connected to the peripheral interface 1103 by buses, signal lines, or circuit boards. Specifically, the peripheral device includes: at least one of radio frequency circuitry 1104, display screen 1105, and audio circuitry 1106.

The Radio Frequency circuit 1104 is used to receive and transmit RF (Radio Frequency) signals, also called electromagnetic signals. The radio frequency circuit 1104 communicates with communication networks and other communication devices via electromagnetic signals.

The display screen 1105 is used to display a UI (user interface). The UI may include graphics, text, icons, video, and any combination thereof. The display screen 1105 may be a touch display screen and may also be used to provide virtual buttons and/or a virtual keyboard.

The audio circuitry 1106 may include a microphone and a speaker. The microphone is used for collecting audio signals of a user and the environment, converting the audio signals into electric signals, and inputting the electric signals to the processor 1101 for processing or inputting the electric signals to the radio frequency circuit 1104 to achieve voice communication. For stereo capture or noise reduction purposes, multiple microphones may be provided, each at a different location of terminal 1100. The microphone may also be an array microphone or an omni-directional pick-up microphone. The speaker is then used to convert the electrical signal from the processor 1101 or the radio frequency circuit 1104 into an audio signal.

Those skilled in the art will appreciate that the configuration shown in fig. 11 is not limiting of terminal 1100, and may include more or fewer components than shown, or some components may be combined, or a different arrangement of components may be used.

Fig. 12 is a schematic structural diagram of a server 1200 according to an embodiment of the present application, where the server 1200 may generate a relatively large difference due to different configurations or performances, and may include one or more processors (CPUs) 1201 and one or more memories 1202, where the memory 1202 stores at least one instruction, and the at least one instruction is loaded and executed by the processors 1201 to implement the methods provided by the foregoing method embodiments. Certainly, the server may further have a wired or wireless network interface, a keyboard, an input/output interface, and other components to facilitate input and output, and the server may further include other components for implementing functions of the device, which are not described herein again.

The server 1200 may be used to perform the above-described information recommendation method or user identification recommendation method.

The embodiment of the present application further provides a computer device, where the computer device includes a processor and a memory, where the memory stores at least one program code, and the at least one program code is loaded and executed by the processor, so as to implement the information recommendation method or the user identification recommendation method of the foregoing embodiment.

The embodiment of the present application further provides a computer-readable storage medium, where at least one program code is stored in the computer-readable storage medium, and the at least one program code is loaded and executed by a processor, so as to implement the information recommendation method or the user identification recommendation method of the foregoing embodiment.

The embodiment of the present application further provides a computer program, where at least one program code is stored in the computer program, and the at least one program code is loaded and executed by a processor, so as to implement the information recommendation method or the user identification recommendation method of the foregoing embodiment.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above description is only an alternative embodiment of the present application and should not be construed as limiting the present application, and any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An information recommendation method, characterized in that the method comprises:

for each second user identifier, selecting a reason generation algorithm according to the probability of each reason generation algorithm in a plurality of reason generation algorithms so as to match the selection times of each reason generation algorithm with the probability, and generating a recommendation reason for the second user identifier by adopting the selected reason generation algorithm;

recommending the user information to the first user identification.

2. The method of claim 1, wherein determining at least one second user identifier to be recommended for the first user identifier comprises:

3. The method of claim 2, wherein said selecting a recommendation algorithm from said plurality of recommendation algorithms comprises:

4. The method of claim 3, wherein after recommending the user information to the first user identification, the method further comprises:

and adjusting the probability of each recommendation algorithm according to the attention degree of the first user identifier to the user information so as to enable the probability of each recommendation algorithm to be matched with the corresponding attention degree.

5. The method of claim 1, wherein determining at least one second user identifier to be recommended for the first user identifier comprises:

6. The method of claim 2 or 5, wherein prior to determining the plurality of recommendation algorithms matching the plurality of user tags based on the plurality of user tags of the first user identification, the method further comprises:

acquiring attribute information of the first user identifier;

7. The method of claim 1, further comprising:

8. The method of claim 1, wherein after recommending the user information to the first user identification, the method further comprises:

and adjusting the probabilities of the multiple reason generation algorithms according to the attention degree of the first user identification to the user information so as to enable the probability of each reason generation algorithm to be matched with the corresponding attention degree.

9. The method of claim 1, wherein generating a recommended reason for the second user identifier using the selected reason generation algorithm comprises:

and when the selected reason generation algorithm is the first reason generation algorithm, selecting a recommendation reason matched with the attribute information from a plurality of fixed recommendation reasons as the recommendation reason of the second user identification according to the attribute information of the second user identification.

10. The method of claim 1, wherein generating a recommended reason for the second subscriber identity using the selected reason generation algorithm comprises:

and when the selected reason generation algorithm is a second reason generation algorithm, adding attribute information of the user identifications associated with the first user identification and the second user identification in a recommendation reason template to generate a common friend recommendation reason.

11. The method of claim 1, wherein generating a recommended reason for the second user identifier using the selected reason generation algorithm comprises:

and when the selected reason generation algorithm is a third reason generation algorithm, generating a recommendation reason for the second user identification based on a text generation model according to the historical recommendation record of each user identification in the category to which the first user identification belongs.

12. The method of claim 11, wherein generating a recommendation reason for the second user identifier based on a text generation model according to the historical recommendation records of each user identifier in the category to which the first user identifier belongs comprises:

according to the historical recommendation record, determining the attention of the recommended user identification in the historical recommendation record to a recommendation reason;

determining a plurality of recommendation reasons according to the attention degree of the recommended user identifier to the recommendation reasons, wherein the attention degree of each recommendation reason in the recommendation reasons is larger than a preset threshold;

and selecting a recommendation reason matched with the attribute information from the plurality of recommendation reasons according to the attribute information of the second user identifier, and determining the recommendation reason as the recommendation reason of the second user identifier.

13. A method for recommending user identification, the method comprising:

14. The method of claim 13, wherein after recommending the at least one second subscriber identity to the first subscriber identity, the method further comprises:

and adjusting the probability of each recommendation algorithm according to the attention degree of the first user identifier to the at least one second user identifier so as to enable the probability of each recommendation algorithm to be matched with the corresponding attention degree.

15. An information recommendation apparatus, characterized in that the apparatus comprises:

the recommendation reason generating module is used for selecting one reason generating algorithm according to the probability of each reason generating algorithm in the multiple reason generating algorithms for each second user identifier so as to enable the selection times of each reason generating algorithm to be matched with the probability, and generating a recommendation reason for the second user identifier by adopting the selected reason generating algorithm;

16. The apparatus of claim 15, wherein the subscriber identity determination module comprises:

the first recommendation algorithm determining unit is used for determining a plurality of recommendation algorithms matched with the user tags according to the user tags of the first user identifier;

17. The apparatus of claim 16, wherein the first subscriber identity determining unit is further configured to select one recommendation algorithm according to the probability of each recommendation algorithm in the plurality of recommendation algorithms, so that the selection times of each recommendation algorithm matches the probability.

18. The apparatus of claim 17, further comprising:

19. The apparatus of claim 15, wherein the subscriber identity determination module comprises:

the second recommendation algorithm determining unit is used for determining a plurality of recommendation algorithms matched with the user tags according to the user tags of the first user identifier;

20. The apparatus of claim 16 or 19, further comprising:

21. The apparatus of claim 15, wherein the referral reason generating module comprises:

22. The apparatus of claim 15, further comprising:

23. The apparatus of claim 15, wherein the recommendation reason generation module comprises:

and the first recommendation reason selecting unit is used for selecting a recommendation reason matched with the attribute information from a plurality of fixed recommendation reasons as the recommendation reason of the second user identification according to the attribute information of the second user identification when the selected reason generating algorithm is the first reason generating algorithm.

24. The apparatus of claim 15, wherein the recommendation reason generation module comprises:

and the second recommendation reason selecting unit is used for adding the attribute information of the user identifications associated with the first user identification and the second user identification in the recommendation reason template to generate a common friend recommendation reason when the selected reason generating algorithm is the second reason generating algorithm.

25. The apparatus of claim 15, wherein the recommendation reason generation module comprises:

and the third reason recommendation selecting unit is used for generating a reason recommendation for the second user identifier based on a text generation model according to the historical recommendation record of each user identifier in the category to which the first user identifier belongs when the selected reason generating algorithm is the third reason generating algorithm.

26. The apparatus according to claim 25, wherein the third reason for recommendation extracting unit is further configured to determine, according to the historical recommendation record, a degree of attention of a recommended user identifier in the historical recommendation record to a reason for recommendation, determine, according to the degree of attention of the recommended user identifier to the reason for recommendation, a plurality of reasons for recommendation, the degree of attention of each reason for recommendation being greater than a preset threshold, extract, according to the attribute information of the second user identifier, a reason for recommendation matching the attribute information from the plurality of reasons for recommendation, and determine the reason for recommendation as the second user identifier.

27. An apparatus for user identification recommendation, the apparatus comprising:

the recommendation algorithm determining module is used for determining a plurality of recommendation algorithms matched with a plurality of user tags according to the plurality of user tags of the first user identifier;

and the user identification recommending module is used for recommending the at least one second user identification to the first user identification.

28. The apparatus of claim 27, further comprising:

29. A computer device, characterized in that the computer device comprises a processor and a memory, wherein at least one program code is stored in the memory, and the at least one program code is loaded and executed by the processor to implement the information recommendation method according to any one of claims 1 to 12; or to implement a user identification recommendation method as claimed in any of claims 13 to 14.

30. A computer-readable storage medium, wherein at least one program code is stored in the computer-readable storage medium, and the at least one program code is loaded and executed by a processor to implement the information recommendation method according to any one of claims 1 to 12; or to implement a user identification recommendation method as claimed in any of claims 13 to 14.