CN107977383A - To the method and device of user's recommending digital content - Google Patents

Abstract

The present invention provides a kind of method and device to user's recommending digital content, one aspect of the present invention obtains user preference vector according to the access behavior of user, on the other hand content adaptation vector can be quantified in turn according to the user characteristics of digital content, when user reuses, required digital content can be recommended to user according to the matching degree between content adaptation vector sum user preference vector, substantially increase the accuracy of digital content recommending.

Description

Method and device for recommending digital content to users

technical field

The invention relates to the field of content recommendation, in particular to a method and device for recommending digital content to users.

Background technique

Digital content refers to information content such as articles, pictures, sounds, and images published on the Internet. With the rapid development of Internet technology, there are more and more types of digital content, how to recommend the digital content to be recommended to interested users has become an important problem in the multimedia field. The traditional digital content recommendation method generally recommends digital content to users based on the attributes of user tags entered manually, and does not consider the adaptation of digital content to users when recommending, and there is a problem of low recommendation accuracy.

Contents of the invention

Embodiments of the present invention propose a method and device for recommending digital content to users, so as to solve the problem of low recommendation accuracy in existing digital content recommendation methods.

In one aspect, an embodiment of the present invention provides a method for recommending digital content to a user, including:

Obtain a user preference vector according to user preference quantification;

Quantify the content adaptation vector according to the user characteristics of the digital content;

Digital content is recommended to the user according to the degree of correlation between the user preference vector and the content adaptation vector.

In another aspect, an embodiment of the present invention provides an apparatus for recommending digital content to a user, including:

A user preference vector determination module, used for user preference quantification to obtain a user preference vector;

A content adaptation vector determination module, configured to quantify and obtain a content adaptation vector according to user characteristics of the digital content;

The content recommendation module is used for recommending digital content to the user according to the degree of correlation between the user preference vector and the content adaptation vector.

The current digital content recommendation method has the problem that the recommendation is not accurate enough and the accuracy is not high. On the one hand, the embodiment of the present invention obtains the user preference vector according to the user's access behavior; Matching vector, when the user uses it again, the required digital content can be recommended to the user according to the matching degree between the content adaptation vector and the user preference vector, which greatly improves the accuracy of digital content recommendation.

Description of drawings

Specific embodiments of the present invention will be described below with reference to the accompanying drawings, wherein:

FIG. 1 shows a schematic flowchart of a method for recommending digital content to a user in an embodiment of the present invention;

Fig. 2 shows the schematic diagram of correcting UPV and CAV in this embodiment;

Fig. 3 A and Fig. 3 B have shown the Ebbinghaus forgetting curve and forgetting factor adopted in the embodiment of the present invention;

Fig. 4 shows a schematic structural diagram of an apparatus for recommending digital content to a user in an embodiment of the present invention.

Detailed ways

In order to make the technical solutions and advantages of the present invention clearer, the exemplary embodiments of the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all implementations. Exhaustive list of examples. And in the case of no conflict, the embodiments in this description and the features in the embodiments can be combined with each other.

The inventor noticed during the invention process that with the rapid development of Internet technology, there are more and more types of digital content, how to recommend digital content to be recommended to interested users has become an important issue in the multimedia field. The traditional digital content recommendation method generally recommends digital content to users based on the attributes of user tags entered manually, and does not consider the adaptation of digital content to users when recommending, and there is a problem of low recommendation accuracy.

In view of the above shortcomings, an embodiment of the present invention proposes a method for recommending digital content to a user, which will be described below.

The implementation of the present invention is based on the following foundations:

(1) Most users can truly provide their own preference information.

(2) The user's content selection preference is related to multiple dimensions (age, gender, occupation, education, hobbies, etc.)

(3) User network access behavior is an objective action based on user preference.

(4) When there are enough samples, the content can also show bias, which reflects which type of users the content is suitable for.

Fig. 1 shows a schematic flowchart of a method for recommending digital content to a user in an embodiment of the present invention, as shown in the figure, including:

Step 101, obtain user preference vectors (UPV, User PreferenceVectors) according to user preference quantification;

Step 102, obtain content adaptation vectors (CAV, ContentAdaptation Vectors) according to the user characteristic quantization of digital content;

Step 103, recommending digital content to the user according to the degree of correlation between the UPV and the CAV.

In this embodiment, on the one hand, the UPV is obtained according to the user's access behavior (such as: watching movies, downloading games), and on the other hand, the CAV can be reversely quantified according to the user characteristics of the digital content. According to the matching degree between them, the required digital content is recommended to users, which greatly improves the accuracy of digital content recommendation.

During specific implementation, in step 101, when the user logs in for the first time, the user needs to fill in the user preference and store it in the user preference database. The filled user preference includes at least one of the following dimensions:

(1) Age (under 20 years old, 20-30 years old, 30-40 years old, 40-50 years old, 50-60 years old, over 60 years old);

(2) Gender (male, female);

(3) Occupation (workers, farmers, private employers, soldiers, students, government and enterprise units, retirees, service personnel);

(4) Educational background (primary and junior high school, high school and technical secondary school, undergraduate and junior college, postgraduate);

(5) Hobbies (literature, sports, reading, travel, etc.).

The initial UPV of the user is determined according to the user preference filled in when the user logs in for the first time. Each dimension of the user preference filled in by the user corresponds to each component of the following UPV. That is, if the user preference filled in by the user when logging in for the first time includes 5 dimensions, then The corresponding UPV includes 5 components, namely:

(1)A(a,b,c,d,e,f)

(2)B(a,b)

(3)C(a,b,c,d,e,f,g,h)

(4)D(a,b,c,d)

(5)E(a,b,c,d)

During specific implementation, in step 102, when the content is provided for the first time, user characteristics for content adaptation need to be filled in, and the format is the same as the user preference filled in by the user in step 101 above, including at least one of the following dimensions:

(1) Age (under 20 years old, 20-30 years old, 30-40 years old, 40-50 years old, 50-60 years old, over 60 years old);

(2) Gender (male, female);

(3) Occupation (workers, farmers, private employers, soldiers, students, government and enterprise units, retirees, service personnel);

(4) Educational background (primary and junior high school, high school and technical secondary school, undergraduate and junior college, postgraduate);

(5) Hobbies (literature, sports, reading, travel, etc.).

The initial CAV is determined according to the content-adapted user characteristics filled in when the content is provided for the first time, and each dimension of the content-adapted user characteristics corresponds to each component of the following CAV one by one, that is, if the filled-in content matches the user The feature includes 5 dimensions, and the corresponding CAV also includes 5 components, which are:

(1)A(a,b,c,d,e,f)

(2)B(a,b)

(3)C(a,b,c,d,e,f,g,h)

(4)D(a,b,c,d)

(5)E(a,b,c,d)

During specific implementation, the process of recommending digital content to the user according to the degree of correlation between UPV and CAV in step 103 is as follows:

Step 1031, calculate the correlation coefficients of the corresponding dimensions between UPV and CAV through a correlation coefficient matching algorithm, such as the Pearson correlation coefficient algorithm, and obtain the correlation coefficient vector |R|(ρ1, ρ2, ρ3, ρ4, ρ5), when |R| is greater than or equal to 0.8, it is considered to be highly correlated; when |R| is greater than or equal to 0.5 and less than 0.8, it is considered to be moderately correlated; relevant;

Step 1032, according to the following table, determine the integral of each component of the correlation coefficient vector of each digital content adaptation user, and sort the results of summing the integrals corresponding to each component of the correlation coefficient vector of each digital content adaptation user;

ρ relativity integral 0.8≤ρ<=1.0 Highly correlated 5 0.5≤ρ<0.8 Moderately relevant 4 0.3≤ρ<0.5 low correlation 2 0≤ρ<0.3 irrelevant 1 -1.0≤ρ≤-0.8 highly negative correlation -5 -0.8≤ρ<-0.5 moderate negative correlation -4 -0.5≤ρ<-0.3 low negative correlation -2 -0.3≤ρ<0 irrelevant -1

Step 1033, obtain the recommended order of digital content according to the sorting result of step 1032, and present digital content to the user according to the order, or intersperse digital content to be promoted by artificially increasing points according to business needs.

Fig. 2 shows a schematic diagram of correcting UPV and CAV in this embodiment.

Furthermore, in this embodiment, according to the computing power of the system, the content will be accessed regularly (for example: 10 minutes, because people's preferences are changeable, it is necessary to ensure that the refresh cycle is small) according to the user's history (note that the content browsed by the user for a short time should be removed, That is, the reading time is lower than the preset threshold, such as 60 seconds), and the UPV is corrected, and the forgetting factor determined according to the Ebbinghaus forgetting model needs to be considered when correcting. Figure 3A and Figure 3B show the The Ebbinghaus forgetting curve and forgetting factor, because the user's recent access behavior can better reflect the user's preference.

The process of correcting the UPV according to the CAV of the user's historical access content is as follows:

Calculate the CAV based on the user's historical access content, and multiply each component of the CAV by the forgetting factor to obtain the arithmetic mean value, that is, the value of each component of the CAV:

Among them: f _n is the forgetting factor obtained by referring to the Ebbinghaus forgetting curve when the user browses the nth item of content; a _n is the first component of the CAV of the user browsing the nth item of content, and the other components are calculated in the same way.

Combine the arithmetic mean value of each component of CAV to obtain CAV', and then obtain CAV" after normalization processing, add the CAV" obtained after normalization processing and the direct vector of UPV, and then normalize each component obtained processing, and finally get the corrected UPV.

Further, in this embodiment, according to the computing power of the system, the UPV (for example, every day) determined based on the user's historical access content (need to exclude the content that the user browses for a short time, that is, the reading time is lower than the preset threshold, such as 60 seconds), correct the CAV, the specific process is as follows:

Calculate the UPV based on the user's historical access content, and take the arithmetic mean of each component of the UPV, namely:

Among them: a _n is the first component of the UPV of the nth user browsed by the user, and the calculation method of other components is the same.

Combine the arithmetic mean value of each component of UPV to obtain UPV', and then obtain UPV" after normalization processing, add the UPV" obtained after normalization processing and the direct vector of CAV, and then normalize each component obtained processing, and finally get the corrected CAV.

In this embodiment, the UPV and CAV are regularly corrected according to the user's historical access content, so that the most accurate recommendation effect can be obtained.

Based on the same inventive concept, an embodiment of the present invention also provides a device for recommending digital content to users. Since the problem-solving principle of the device in this embodiment is similar to the above-mentioned method for recommending digital content to users, this embodiment For the implementation of the example, refer to the implementation of the method, and the repetition will not be repeated.

Fig. 4 shows a schematic structural diagram of a device for recommending digital content to users in an embodiment of the present invention. As shown in the figure, the device may include:

The user preference vector determination module 401 obtains the UPV according to the user preference quantization;

A content adaptation vector determination module 402, configured to quantify and obtain CAV according to user characteristics of digital content;

The content recommendation module 403 is configured to recommend digital content to the user according to the degree of correlation between the UPV and the CAV.

The dimensions of UPV include at least one of age, gender, occupation, education and hobbies;

The dimensions of CAV include at least one of age, gender, occupation, education and hobbies.

The content recommendation module 403 calculates the correlation coefficient of the corresponding dimension between UPV and CAV through the correlation coefficient matching algorithm, obtains the correlation coefficient vector of each digital content adaptation user, and according to the integral sum of each dimension of the correlation coefficient vector of each digital content adaptation user Sort and recommend digital content to users.

Further, the user preference vector determination module 401 is also used to periodically revise the UPV according to the CAV of the user's historical access content, and the revision process is as follows:

Calculate the CAV based on the user's historical access content, multiply each component of the CAV by the forgetting factor and correct it to take the arithmetic mean value, and the forgetting factor is obtained according to the Ebbinghaus forgetting curve;

Combine the arithmetic mean value of each component of CAV to obtain CAV', and then obtain CAV" after normalization processing, add the CAV" obtained after normalization processing and the direct vector of UPV, and then normalize each component obtained processing, and finally get the corrected UPV.

Further, the content adaptation vector determination module 402 is also used to periodically correct the CAV according to the UPV determined by the user's historical access content, and the correction process is as follows:

Calculate the UPV based on the user's historical access content, and take the arithmetic mean of each component of the UPV;

Combine the arithmetic mean value of each component of UPV to obtain UPV', and then obtain UPV" after normalization processing, add the UPV" obtained after normalization processing and the direct vector of CAV, and then normalize each component obtained processing, and finally get the corrected CAV.

For the convenience of description, each part of the device described above is divided into various modules or units by function and described separately. Of course, when implementing the present invention, the functions of each module or unit can be implemented in one or more pieces of software or hardware.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, systems, or computer program products. Accordingly, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

While preferred embodiments of the invention have been described, additional changes and modifications to these embodiments can be made by those skilled in the art once the basic inventive concept is appreciated. Therefore, it is intended that the appended claims be construed to cover the preferred embodiment as well as all changes and modifications which fall within the scope of the invention.

Claims (11)

1. A method for recommending digital content to users, comprising: Obtain a user preference vector according to user preference quantification; Quantify the content adaptation vector according to the user characteristics of the digital content; Digital content is recommended to the user according to the degree of correlation between the user preference vector and the content adaptation vector. 2. The method according to claim 1, wherein the dimension of the user preference vector includes at least one of age, gender, occupation, education and hobbies; The dimensions of the content adaptation vector include at least one of age, gender, occupation, education and hobbies. 3. The method according to claim 2, wherein said recommending digital content to the user according to the degree of correlation between the user preference vector and the content adaptation vector comprises: Calculate the correlation coefficient of the corresponding dimension between the content adaptation vector and the user preference vector through the correlation coefficient matching algorithm, and obtain the correlation coefficient vector of each digital content adaptation user; Determine the ranking of the integral sum of each component of the correlation coefficient vector for each digital content adapted to the user, and recommend digital content to the user. 4. The method according to claim 1 or 2, further comprising the step of regularly revising the user preference vector according to the content adaptation vector of the user's historical access content. 5. The method according to claim 4, wherein the process of modifying the user preference vector according to the content adaptation vector of the user's historical access content is as follows: Calculate the content adaptation vector according to the user's historical access content, multiply each component of the content adaptation vector by the forgetting factor and correct and take the arithmetic mean value, and the forgetting factor is obtained according to the Ebbinghaus forgetting curve; The arithmetic mean value of each component of the content adaptation vector is combined and normalized, and the user preference vector is corrected by using the normalized content adaptation vector. 6. The method according to claim 1 or 2, further comprising the step of regularly revising the content adaptation vector based on the user preference vector determined from the user's historical content access. 7. The method according to claim 6, wherein the process of modifying the content adaptation vector is as follows: Calculate the user preference vector according to the user's historical access content, and take the arithmetic mean value of each component of the user preference vector; The arithmetic mean value of each component of the user preference vector is combined and normalized, and the content adaptation vector is corrected by using the normalized user preference vector. 8. A device for recommending digital content to users, comprising: A user preference vector determination module, used for user preference quantification to obtain a user preference vector; A content adaptation vector determination module, configured to quantify and obtain a content adaptation vector according to user characteristics of the digital content; The content recommendation module is used for recommending digital content to the user according to the degree of correlation between the user preference vector and the content adaptation vector. 9. The device according to claim 8, wherein the dimension of the user preference vector includes at least one of age, gender, occupation, education and hobbies; The dimensions of the content adaptation vector include at least one of age, gender, occupation, education and hobbies. 10. The device according to claim 8, wherein the content recommendation module calculates the correlation coefficient of the corresponding dimension between the content adaptation vector and the user preference vector through a correlation coefficient matching algorithm, and obtains each digital content adaptation user's The correlation coefficient vector is used to recommend the digital content to the user according to the ranking of the integral sum of each component of the correlation coefficient vector of the user for each digital content. 11. The device according to claim 8, wherein the user preference vector determining module is further configured to periodically revise the user preference vector according to the content adaptation vector of the user's historical access content; The content adaptation vector determination module is further configured to periodically revise the content adaptation vector based on the user preference vector determined from the user's historical access to content.