CN111291217B - Content recommendation method, device, electronic equipment and computer readable medium - Google Patents
Content recommendation method, device, electronic equipment and computer readable medium Download PDFInfo
- Publication number
- CN111291217B CN111291217B CN202010144164.7A CN202010144164A CN111291217B CN 111291217 B CN111291217 B CN 111291217B CN 202010144164 A CN202010144164 A CN 202010144164A CN 111291217 B CN111291217 B CN 111291217B
- Authority
- CN
- China
- Prior art keywords
- matrix
- recommended content
- interest
- degree
- content
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 239000011159 matrix material Substances 0.000 claims abstract description 181
- 238000004891 communication Methods 0.000 claims description 19
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 8
- 238000000354 decomposition reaction Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 description 9
- 239000013598 vector Substances 0.000 description 9
- 230000009471 action Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/73—Querying
- G06F16/735—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
A content recommendation method, device, electronic equipment and computer readable medium are disclosed, which are implemented by acquiring a history click record of media content of interest to a user; calculating the degree of interest in the media content by the user based on the history click record; calculating a customized recommended content matrix according to the interest extensive degree of the user and a preset standard recommended content matrix, wherein the value of each element in the customized recommended content matrix is used for representing the recommendation priority of the standard recommended content; based on the recommendation priority, selecting recommended content from a plurality of standard recommended content, calculating personalized interest extensive degree aiming at different interest extensive degrees of users on media content, then combining the interest extensive degree with a preset standard recommended content matrix to calculate the priority of the standard recommended content, and finally selecting recommended content meeting the user interest extensive degree from the plurality of standard recommended content based on the priority, thereby realizing flexible personalized recommendation and improving user experience.
Description
Technical Field
The present disclosure relates to the field of communications, and in particular, to a content recommendation method, apparatus, electronic device, and computer readable medium.
Background
With the advent of the big data age, people are filled with different types of recommended contents in daily life, and platforms of different industries such as application software, video platforms, shopping platforms and the like are all in great charge of recommending contents to users so as to mine new interest points of the users. For example: in the video platform, the video recommendation system can recommend the same type of video according to the historical browsing of the user, and then recommend the related type of video in order to meet the diversity requirement of the user.
However, the interest ranges of different users are different, for example, some users have wide interests, like various types of related videos pushed by a recommendation system, some users have single interests, and only like the same type of videos pushed by the recommendation system, however, the current recommendation system does not consider the different interest ranges of different users, uniformly recommends the pre-selected videos, and causes dead recommendation content and poor user experience.
Disclosure of Invention
The embodiment of the invention aims to provide a content recommendation method, a device, electronic equipment and a computer readable medium, so as to solve the technical problems that a current recommendation system does not consider different interest ranges of different users, and uniformly recommends preselected videos, so that recommended content is dead and poor in user experience. The specific technical scheme is as follows:
In a first aspect of the application, there is provided a content recommendation method, including:
if a content recommendation request of a user is received, acquiring a historical click record of the user;
calculating the degree of interest in the media content by the user based on the historical click record;
calculating a customized recommended content matrix according to the user interest extensive degree and a preset standard recommended content matrix, wherein the standard recommended content matrix and elements at the same position in the customized recommended content matrix correspond to the same standard recommended content, and the value of each element in the customized recommended content matrix is used for representing the recommended priority of the standard recommended content;
and selecting recommended content from a plurality of standard recommended content based on the recommended priority.
Optionally, the step of calculating the popularity of the user's interest in the media content based on the history of clicks comprises:
determining the category number of the media content interested by the user and the attention degree corresponding to each type according to the historical click record;
and calculating the interest degree according to the category number and the attention degree corresponding to each type.
Optionally, a formula for calculating the degree of interest according to the number of categories and the degree of interest corresponding to each type is as follows:
wherein W is u Represents the degree of interest, M represents the number of species, N ui Represent the degree of interest of type i, N u Representing the sum of the attention levels corresponding to each type.
Optionally, the step of calculating the customized recommended content matrix according to the user's interest degree and the preset standard recommended content matrix includes:
reversely decomposing the standard recommended content matrix to obtain a correlation matrix and a similarity matrix;
calculating the product of the interest extensive degree and the similarity matrix to obtain a customized similarity matrix;
and generating the customized recommended content matrix according to the customized similarity matrix and the correlation matrix.
In a second aspect of the present application, there is also provided a content recommendation apparatus, including:
the acquisition module is used for acquiring a historical click record of the user if a content recommendation request of the user is received;
a first calculation module for calculating the degree of interest in the media content by the user based on the history of clicks;
the second calculation module is used for calculating a customized recommended content matrix according to the interest extensive degree of the user and a preset standard recommended content matrix, wherein the standard recommended content matrix and the same position element in the customized recommended content matrix correspond to the same standard recommended content, and the value of each element in the customized recommended content matrix is used for representing the recommended priority of the standard recommended content;
And the selecting module is used for selecting recommended content from a plurality of standard recommended content based on the recommended priority.
Optionally, the first computing module includes:
a determining unit, configured to determine the number of categories of media content interested by the user and a degree of attention corresponding to each type according to the history click record;
a first calculation unit for calculating the degree of interest spread according to the number of categories and the degree of interest corresponding to each type.
Optionally, a formula for calculating the degree of interest according to the number of categories and the degree of interest corresponding to each type is as follows:
wherein W is u Represents the degree of interest, M represents the number of species, N ui Represent the degree of interest of type i, N u Representing the sum of the attention levels corresponding to each type.
Optionally, the second computing module includes:
the decomposition unit is used for reversely decomposing the standard recommended content matrix to obtain a correlation matrix and a similarity matrix;
the second calculation unit is used for calculating the product of the interest extensive degree and the similarity matrix to obtain a customized similarity matrix;
and the generation unit is used for generating the customized recommended content matrix according to the customized similarity matrix and the correlation matrix.
In a third aspect of the present application, there is further provided an electronic device, including a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any of the above first aspects when executing a program stored on a memory.
In a fourth aspect of the present application, there is also provided a computer readable storage medium having stored thereon a computer program, characterized in that the program, when executed by a processor, implements the method of any of the above first aspects.
Compared with the prior art, the technical scheme provided by the embodiment of the application has the following advantages:
if a content recommendation request of a user is received, acquiring a historical click record of the user; calculating the degree of interest in the media content by the user based on the historical click record; calculating a customized recommended content matrix according to the user interest extensive degree and a preset standard recommended content matrix, wherein the standard recommended content matrix and elements at the same position in the customized recommended content matrix correspond to the same standard recommended content, and the value of each element in the customized recommended content matrix is used for representing the recommended priority of the standard recommended content; and selecting recommended content from a plurality of standard recommended content based on the recommended priority.
According to the method, firstly, the historical click record of the user is obtained, the interest extensive degree is calculated based on the historical click record, the personalized interest extensive degree is calculated according to different interest extensive degrees of the user on media content, the interest extensive degree is combined with a preset standard recommended content matrix to calculate the priority of standard recommended content, and finally recommended content which accords with the user interest extensive degree is selected based on a plurality of priority standard recommended content, so that flexible personalized recommendation is realized, and user experience is improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, and it will be obvious to a person skilled in the art that other drawings can be obtained from these drawings without inventive effort.
Fig. 1 is a flow chart of a content recommendation method according to an embodiment of the present application;
Fig. 2 is a schematic flowchart of step S102 in fig. 1 according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a standard recommendation content matrix establishment procedure according to an embodiment of the present application;
fig. 4 is a schematic flowchart of step S103 in fig. 1 provided in an embodiment of the present application;
fig. 5 is a schematic flowchart of step S104 in fig. 1 provided in an embodiment of the present application;
FIG. 6 is a block diagram of a content recommendation device according to an embodiment of the present invention;
fig. 7 is a block diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present application based on the embodiments herein.
At present, people are filled with different types of recommended contents in daily life, platforms of different industries such as application software, video platforms and shopping platforms are all in great effort to recommend contents to users so as to mine new interest points of the users, most of multimedia platforms recommend some preset media contents to the users, however, different users have different interest ranges, for example, the interests of the users are wide, various types of related videos (life, military, science and technology and the like) which are favored by a recommendation system are single, and the users are only interested in the same type of videos which are favored by the recommendation system, so that the current recommendation system uniformly recommends the preselected videos to the users, the recommended contents die, and the recommended contents do not fit the requirements of the users, thereby resulting in poor user experience. Based on this, the content recommendation method provided in the embodiment of the present application, as shown in fig. 1, includes:
Step S101, if a content recommendation request of a user is received, acquiring a historical click record of the user;
in the embodiment of the present application, the content recommendation request may be a request for acquiring recommended content, and in an actual application, whether the recommended content accords with the interest of the user directly feels the quality of the recommended content, so that in the embodiment of the present application, the historical click record of the user is firstly acquired to determine the interest direction of the user.
The manner of acquiring the historical click record of the user may depend on the actual situation, for example: in the practical application of the video website platform, after a user clicks a video, the video classification label of the clicked video of the user is recorded in the background and stored in the background so as to analyze the media content of interest to the user. The above-described embodiments are merely examples, and the embodiments of the present application are not particularly limited thereto.
Step S102, calculating the degree of interest of the user in the media content based on the historical click record;
in the embodiments of the present application, the media content refers to a medium for transmitting information, which refers to a means, a channel, a carrier, an intermediary or a technical means for transmitting information and acquiring information, and may also be regarded as all technical means for implementing information transmission from an information source to a recipient, for example: audio, video, and other forms of communicating information. The degree of popularity of interest indicates a tendency of a user to media content, for example, media content representing different subjects (military subjects, living subjects, scientific subjects, fun subjects, variety subjects, etc.) liked by the user is high in the degree of popularity of the user, while media content representing the liked by the user is relatively single in the degree of popularity of the user.
In this step, the manner of calculating the interest extensive degree through the history click record may depend on the actual situation, and in an exemplary manner, the history click record includes the user history browsing videos and classification labels corresponding to the videos, and the probability of the user browsing videos of each category may be calculated through the probability statistics manner, so as to determine whether the user browsing videos are of the same category or of various categories, and further determine the interest extensive degree of the user for the media content.
Step S103, calculating a customized recommended content matrix according to the user interest extensive degree and a preset standard recommended content matrix, wherein the standard recommended content matrix and elements at the same position in the customized recommended content matrix correspond to the same standard recommended content, and the value of each element in the customized recommended content matrix is used for representing the recommended priority of the standard recommended content;
in the embodiment of the present application, the standard recommended content may be preset manually, and the present application is used to characterize the recommendation priority of each standard recommended content by establishing a standard recommended content matrix, where the recommendation priority may be preset manually.
In practical applications, for example: a video platform selects 10 top-quality videos as recommended contents for users, and establishes a standard recommended content matrix for representing the recommended priority of the recommended contents, wherein a plurality of elements are used for representing the recommended priority of the standard recommended contents in the matrix, and each element in the matrix is used for representing the recommended priority of the standard recommended contents, so that whether the recommended contents meet the recommended condition can be determined by calculating the recommended priority of the plurality of elements used for representing the recommended contents of the standard.
In the step, the mode of calculating the customized recommended content matrix according to the interest extensive degree and the preset standard recommended content matrix can be determined according to actual conditions, the interest extensive degree is introduced into the standard recommended content matrix according to a preset algorithm, and the customized recommended content matrix can be obtained.
And step S104, selecting recommended content from a plurality of standard recommended content based on the recommended priority.
In the embodiment of the application, the recommended content is screened out from a plurality of standard recommended content and is used for being pushed to a user, the method calculates the extensive degree of interest of the user on the media content through historical information such as historical browsing records of the user, then combines the extensive degree of interest with standard content recommendation matrixes representing a plurality of standard recommended content to obtain a customized recommended content matrix representing the recommended priority of the recommended content, so that the priority of each standard recommended content is calculated according to the extensive degree of interest of the user, and finally the recommended content which accords with the extensive degree of interest of the user can be selected from a plurality of standard recommended content based on the recommended priority, the problem of single dead plate of the content pushed by the traditional recommendation system is solved, the recommended content is enabled to be more in accordance with user preference, and the user experience is improved.
In yet another embodiment of the present application, a specific implementation of calculating the interest level through a history of click records is provided, as shown in fig. 2, step S102, calculating the interest level of the user in the media content based on the history of click records includes:
step S201, determining the category number of the media content interested by the user and the attention degree corresponding to each type according to the history click record;
in this embodiment of the present application, the number of categories of media content may be the number of class labels corresponding to media content browsed by a user, in practical application, each media content may be corresponding to a class one class label, where the class one class label is used to represent a maximum partition of media content, for example, a military topic, a variety of topics, a laughing topic, etc., each time a user browses a media content, the background will be recorded, the more the number of times the user browses a topic a, the more the user pays attention to the topic, the fewer the number of times the user browses a topic B, the less the user pays attention to the topic, and the history click record may be used to represent a history browsing record of the user, so the number of categories of media content of interest to the user and the attention degree corresponding to each type are extracted from the history click record.
Step S202, calculating the interest extensive degree according to the category number and the attention degree corresponding to each type.
In the embodiment of the application, the degree of interest of the user can be calculated by counting the types of the media content of interest of the user and the attention degree corresponding to each type, for example: according to the historical click record, the attention degree of the user to the A type content is determined to be 25%, the attention degree to the B type content is determined to be 15%, the attention degree to the C type content is determined to be 35%, the attention degree to the D type content is determined to be 25%, and the interest degree can be calculated according to a preset statistical formula, further, in step S202, the formula for calculating the interest degree according to the category number and the attention degree corresponding to each type is as follows:
wherein W is u Represents the degree of interest, M represents the number of species, N ui Represent the degree of interest of type i, N u Representing the sum of the attention levels corresponding to each type.
In the embodiment of the application, the category number and the attention degree corresponding to each type are brought into the formula established based on the information entropy concept, so that the interest degree can be calculated, and the information has uncertainty, namely the uncertainty exists in the probability that the user browses each type of media content, so that the interest degree calculation formula established based on the information entropy concept only focuses on the possibility of the user browsing the media content to a wide extent, namely the interest degree, and the obtained W u Ranging from 0 to 1.
According to the method and the device for recommending the media content, the interest range of the user to the media content of interest is determined by calculating the interest extensive degree of the user, and media content fitting the interest can be recommended for the user according to different interests of different users.
In still another embodiment provided in the present application, for the method for establishing a standard physical examination content matrix in a media content recommendation system, optionally, as shown in fig. 3, the process for establishing a standard recommendation content matrix includes:
step S301, obtaining a plurality of standard recommended contents and feature vectors corresponding to each standard recommended content;
in this step, the feature vector corresponding to the standard recommended content is a vector representing the content attribute of the standard recommended content, and in practical application, the standard recommended content may be a recommended video preset by a video website, where each video corresponds to a feature vector representing the topic type of the video, for example: the video design is a funneling element and a variety element, and in the subject type space, the feature vector of the video is biased to the funneling element and the variety element.
Step S302, calculating a relevance score for representing the relevance degree between the standard recommended content and the historical click record according to the standard recommended content and the historical click record for each standard recommended content, and obtaining a plurality of relevance scores;
In this step, the standard recommended content is preset, and the history clicking record is obtained according to the media content interested by the user, so that a relevance score can be calculated through the standard recommended content and the history clicking record, the relevance degree can be the type relevance degree between the standard recommended content and the history clicking record, the specific calculation mode can be determined according to the actual situation, and any calculation mode capable of calculating the relevance degree between the history clicking record and the standard recommended content can be selected, which is not limited in the embodiment of the present application.
Step S303, sorting the plurality of relevance scores according to the sequence from the big to the small of the relevance scores to obtain a standard recommended content sequence;
in this step, after calculating the relevance score between each standard recommended content and the history of clicks, a plurality of the standard recommended content are ordered according to the order of the relevance scores from the large to the small, for example: and the relevance score of the video A is 90, the relevance score of the video B is 68 and the relevance score of the video C is 88, and the three standard recommended contents are sequenced according to the sequence from the big to the small of the relevance score to obtain {90, 88, 68}, wherein each relevance score corresponds to one video number.
Step S304, generating a correlation matrix according to the standard recommended content sequence;
in this step, since the correlation matrix is generated by the standard recommended content sequence, the correlation matrix may represent a degree of correlation between the standard recommended content and the historical clicks of the media content of interest to the user.
In this step, the specific operation manner of generating the correlation matrix according to the standard recommended content sequence may be determined according to practical situations, which is not specifically limited in the embodiment of the present application.
Step S305, calculating the product of the feature vectors of any two standard recommended contents to obtain a plurality of similarity scores for representing the similarity degree of the two standard recommended contents;
in this step, the feature vector corresponding to each standard recommended content is used to represent the vector of the content attribute of the standard recommended content, so the product of the feature vectors of any two standard recommended contents is calculated, and the obtained similarity score is used to represent the similarity degree between the two standard recommended contents, wherein the two standard recommended contents selected arbitrarily can be different or the same.
Step S306, generating a similarity matrix according to a plurality of similarity scores;
In this step, since the similarity matrix is generated from a plurality of the similarity scores, the similarity matrix may represent the degree of similarity between any two standard recommended contents. The specific operation manner of generating the similarity matrix according to the plurality of similarity scores may be determined according to practical situations, and the embodiment of the present application is not limited in particular.
Step S307, generating the standard recommended content matrix according to the correlation matrix and the similarity matrix;
in this step, the standard recommended content may be manually preset, a correlation matrix for representing the degree of correlation between the standard recommended content and the historical click record of the media content of interest to the user may be calculated according to the historical click record of interest to the user and the standard recommended content, that is, the degree of interest of the user to the standard recommended content may be expressed, then a similarity matrix for representing the degree of correlation between any two standard recommended contents may be calculated, and finally the standard recommended content matrix may be generated according to the correlation matrix and the similarity matrix, so as to achieve the purpose of calculating a recommendation priority for each standard recommended content, and further the standard recommended content meeting the interest of the user may be selected.
Based on the established standard content recommendation matrix, the present application can calculate a customized recommended content matrix matching the interests of the user by combining the interests of the user, and in another embodiment provided in the present application, a specific embodiment of calculating the customized recommended content matrix according to the interests of the user and the preset standard recommended content matrix is provided, as shown in fig. 4, step S103, includes the steps of:
s401, reversely decomposing the standard recommended content matrix to obtain a correlation matrix and a similarity matrix;
in the embodiment of the present application, in order to introduce a pre-calculated user interest into a standard recommended content matrix to a wide extent, it is necessary to first perform matrix inversion decomposition on the standard recommended content matrix to obtain a correlation matrix and a similarity matrix, and the specific manner in which the standard recommended content matrix is formed by the correlation matrix and the similarity matrix may refer to the content of the foregoing embodiment, and similarly, the standard recommended content matrix may be subjected to matrix inversion decomposition to obtain the correlation matrix and the similarity matrix, which will not be repeated herein.
Step S402, calculating the product of the interest extensive degree and the similarity matrix to obtain a customized similarity matrix;
in the embodiment of the application, the interest extensive degree of the user is taken as a weight to be introduced into the similarity matrix by calculating the product of the interest extensive degree and the similarity matrix to obtain the customized similarity matrix, so that the aim of adjusting the priority corresponding to the standard recommended content according to the interest extensive degree of the user is fulfilled, wherein when the interest extensive degree W u Approaching 1, the media content types of interest to the user are shown to be relatively broad, that is, the user wants to see more recommended content of different types, so the interest is broad degree W u When the similarity matrix is close to 1, the similarity matrix is also close to the original matrix, and the media content recommended to the user is basically consistent with the original preset; when the interest is wide degree W u Near 0, it is shown that the user's interests are more concentrated and the similarity preferences are weaker, i.e., the customized similarity matrix has little effect, and the final recommended content types are more concentrated.
Step S403, generating the customized recommended content matrix according to the customized similarity matrix and the correlation matrix;
In the embodiment of the present application, the specific implementation manner of generating the customized recommended content matrix according to the customized similarity matrix and the relevance matrix may be determined according to practical situations, for example: optionally, in an embodiment of the present application, a specific implementation manner of generating a customized recommended content matrix according to a customized similarity matrix and a correlation matrix is provided, and step S402 includes the following formula of generating the customized recommended content matrix according to the customized similarity matrix and the correlation matrix:
θ∈[0,1]
wherein S represents the customized similarity matrix,represents the correlation matrix, θ represents the adjustment parameter between the preset correlation and similarity, R C Representing the relevance score, W u Indicating the degree of interest, L N×N Representing elements in the customized recommended content matrix.
The θ may be a preset value for adjusting a relationship between the correlation and the similarity of the standard recommended content, the smaller θ is, the more similar the recommended result is, the more abundant and diversified the media content recommended to the user is, the larger θ is, the more the recommended result is biased to the correlation, the type of the media content recommended to the user is single, and the specific value of θ may be according to the actual situation, which is not particularly limited in the embodiment of the present application.
In the embodiment of the application, a DPP (Determinantal Point Process, determinant point process) diversity control algorithm is adopted to establish a customized recommended content matrix, and the embodiment of the application achieves the aim of calculating the priority of standard recommended content by combining the extensive degree of interest of the user by introducing the extensive degree of interest of the user into the customized similarity matrix S.
In practical application, when the type of the media content of interest of the user is relatively wide, the calculated interest degree W u Then it approaches 1, at which time the L matrix approximates the original L matrix, and no change occurs, i.e., at which time the media content recommended to the user is substantially consistent with the original preset; when the media content types of interest of the user are single, the calculated interest degree W u Near 0, it is shown that the user's interests are more concentrated and the similarity preferences are weaker, i.e., the customized similarity matrix has little effect, and the final recommended content types are more concentrated.
In the embodiment provided in the present application, the solving manner of the customized recommended content matrix established by the DPP diversity control algorithm, that is, the determining manner of the recommended content, optionally, as shown in fig. 5, step S104, based on the recommendation priority, selects the recommended content from a plurality of standard recommended contents, includes:
Step S501, randomly acquiring K from the customized recommended content matrix 2 Constructing a submatrix by the individual elements to obtain a plurality of submatrices, wherein K is the number of preset recommended contents;
in this embodiment of the present application, N is a preset number of recommended content, for example, a video website wants to recommend 3 video contents to a user, where N is 3, 9 elements are randomly acquired from a customized recommended content matrix, preferably, 9 elements with high association degree are selected from the customized recommended content matrix, that is, elements with the same row number and column number are selected to represent a recommendation score of a certain video, for example: selecting the elements of the first row and the first column to construct a submatrix A, and then recommending contents represented by the submatrix A and R c1 The corresponding standard recommendations are identical.
In one practical application, setting N to 3, randomly acquiring 9 elements from the customized recommended content matrix, and selecting elements in 1 st, 3 rd and 5 th rows and in 1 st, 3 rd and 5 th columns to obtain L 11 、L 13 、L 15 、L 31 、L 33 、L 35 、L 51 、L 53 、L 55 And constructing a submatrix by using the elements to obtain a matrix for representing R c1 Corresponding standard recommended content, R c2 Corresponding standard recommended content and R c3 A sub-matrix of recommendation scores for corresponding standard recommended content.
Step S502, calculating determinant values of each submatrix according to recommended priorities corresponding to each element in the submatrix for each submatrix;
In the embodiment of the application, the recommendation score is calculated on the randomly extracted standard recommendation content by adopting a mode of calculating the determinant value of the submatrix so as to determine the optimal recommended scheme.
Step S503, determining the submatrix with the largest determinant value as the recommended content matrix;
in this embodiment of the present application, if the customized recommendation matrix is an M-order matrix, N recommendation contents are selected from the M-order matrix, where there areIn one case, the embodiment of the present application passes the formula:
calculating to obtain a submatrix with the maximum determinant value, and taking the submatrix as a recommended content matrix, wherein Y is a subset which belongs to a set C and has a length N (the number of the finally deduced videos is 10, for example); l (L) Y Is a submatrix composed of rows and columns corresponding to Y in the L matrix (e.g., y= {1,2}, ly= [ L11, L12; L21, L22)])；det(L Y ) Represents L Y Is a determinant value of (a).
And step S504, determining the recommended content according to standard recommended content corresponding to the elements of the K recommended content matrixes.
In the embodiment of the application, illustratively, the elements in rows 1, 3 and 5 and in columns 1, 3 and 5 are selected to obtain L 11 、L 13 、L 15 、L 31 、L 33 、L 35 、L 51 、L 53 、L 55 And constructing a submatrix using the elements, L is defined as L because the standard recommended content sequence orders the plurality of relevance scores in order of the relevance scores from large to small 11 、L 13 、L 15 、、L 31 、L 51 For characterizing video 1 in set C, and so on, L 13 、L 31 、L 33 、L 35 、L 53 For representing video 3 in set C, L 15 、L 35 、L 51 、L 53 、L 55 Representing the 5 th video in set C.
According to the embodiment of the application, based on the DPP algorithm solving principle, the solution of the customized standard matrix is realized, and the determinant of the submatrix is calculated based on each element used for representing the video recommendation priority in the matrix, so that the optimal solution is solved from the customized recommendation matrix constructed based on the user interest extensive degree, namely the recommendation content of the user interest is attached, and the effect of improving the user experience is achieved.
In still another embodiment of the present application, there is further provided a content recommendation apparatus, as shown in fig. 6, including:
the acquisition module 01 is used for acquiring a historical click record of the media content of interest to the user;
a first calculation module 02, configured to calculate, based on the history of clicks, how broadly the user is interested in the media content;
a second calculation module 03, configured to calculate a customized recommended content matrix according to the user's interest popularity and a preset standard recommended content matrix, where the standard recommended content matrix and elements at the same position in the customized recommended content matrix correspond to the same standard recommended content, and a value of each element in the customized recommended content matrix is used to represent a recommendation priority of the standard recommended content;
And the selecting module 04 is used for selecting recommended content from a plurality of standard recommended content based on the recommended priority.
In yet another embodiment provided herein, the first computing module includes:
a determining unit, configured to determine the number of categories of media content interested by the user and a degree of attention corresponding to each type according to the history click record;
a first calculation unit for calculating the degree of interest spread according to the number of categories and the degree of interest corresponding to each type.
Optionally, a formula for calculating the degree of interest according to the number of categories and the degree of interest corresponding to each type is as follows:
wherein W is u Represents the degree of interest, M represents the number of species, N ui Represent the degree of interest of type i, N u Representing the sum of the attention levels corresponding to each type.
In yet another embodiment provided herein, the second computing module includes:
the decomposition unit is used for reversely decomposing the standard recommended content matrix to obtain a correlation matrix and a similarity matrix;
the second calculation unit is used for calculating the product of the interest extensive degree and the similarity matrix to obtain a customized similarity matrix;
and the generation unit is used for generating the customized recommended content matrix according to the customized similarity matrix and the correlation matrix.
The embodiment of the present invention also provides an electronic device, as shown in fig. 7, including a processor 1110, a communication interface 1120, a memory 1130, and a communication bus 1140, where the processor 1110, the communication interface 1120, and the memory 1130 perform communication with each other through the communication bus 1140,
a memory 1130 for storing a computer program;
processor 1110, when executing the program stored in memory 1130, performs the following steps:
acquiring a historical click record of media content of interest to a user; calculating the degree of interest in the media content by the user based on the historical click record; calculating a customized recommended content matrix according to the user interest extensive degree and a preset standard recommended content matrix, wherein the standard recommended content matrix and elements at the same position in the customized recommended content matrix correspond to the same standard recommended content, and the value of each element in the customized recommended content matrix is used for representing the recommended priority of the standard recommended content; and selecting recommended content from a plurality of standard recommended content based on the recommended priority.
The communication bus mentioned by the above terminal may be a peripheral component interconnect standard (Peripheral Component Interconnect, abbreviated as PCI) bus or an extended industry standard architecture (Extended Industry Standard Architecture, abbreviated as EISA) bus, etc. The communication bus may be classified as an address bus, a data bus, a control bus, or the like. For ease of illustration, the figures are shown with only one bold line, but not with only one bus or one type of bus.
The communication interface is used for communication between the terminal and other devices.
The memory may include random access memory (Random Access Memory, RAM) or non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory may also be at least one memory device located remotely from the aforementioned processor.
The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU for short), a network processor (Network Processor, NP for short), etc.; but also digital signal processors (Digital Signal Processing, DSP for short), application specific integrated circuits (Application Specific Integrated Circuit, ASIC for short), field-programmable gate arrays (Field-Programmable Gate Array, FPGA for short) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
In yet another embodiment of the present invention, a computer readable storage medium is provided, in which instructions are stored, which when run on a computer, cause the computer to perform the content recommendation method according to any of the above embodiments.
In a further embodiment of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the content recommendation method of any of the above embodiments.
In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the present invention, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted from one computer-readable storage medium to another, for example, by wired (e.g., coaxial cable, optical fiber, digital Subscriber Line (DSL)), or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), etc.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
In this specification, each embodiment is described in a related manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.
The foregoing description is only of the preferred embodiments of the present invention and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.
It should be noted that in this document, relational terms such as "first" and "second" and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (6)
1. A content recommendation method, the method comprising:
if a content recommendation request of a user is received, acquiring a historical click record of the user;
calculating the user's popularity of interest in media content based on the history of clicks, comprising: determining the category number of the media content interested by the user and the attention degree corresponding to each type according to the historical click record; calculating the interest extensive degree according to the category number and the attention degree corresponding to each type;
the formula for calculating the interest popularity according to the category number and the attention degree corresponding to each type is as follows:
Wherein W is u Represents the degree of interest, M represents the number of species, N ui Represent the degree of interest of type i, N u Representing a sum of the attention degrees corresponding to each type;
calculating a customized recommended content matrix according to the user interest extensive degree and a preset standard recommended content matrix, wherein the standard recommended content matrix and elements at the same position in the customized recommended content matrix correspond to the same standard recommended content, and the value of each element in the customized recommended content matrix is used for representing the recommended priority of the standard recommended content;
and selecting recommended content from a plurality of standard recommended content based on the recommended priority.
2. The content recommendation method according to claim 1, wherein the step of calculating a customized recommended content matrix according to the user's interest popularity and a preset standard recommended content matrix comprises:
reversely decomposing the standard recommended content matrix to obtain a correlation matrix and a similarity matrix;
calculating the product of the interest extensive degree and the similarity matrix to obtain a customized similarity matrix;
and generating the customized recommended content matrix according to the customized similarity matrix and the correlation matrix.
3. A content recommendation device, comprising:
the acquisition module is used for acquiring a historical click record of the user if a content recommendation request of the user is received;
a first calculation module for calculating the degree of interest in the media content by the user based on the history of clicks;
the first computing module includes: a determining unit, configured to determine the number of categories of media content interested by the user and a degree of attention corresponding to each type according to the history click record; a first calculation unit configured to calculate the degree of interest spread from the number of categories and the degree of interest corresponding to each type; the formula for calculating the interest popularity according to the category number and the attention degree corresponding to each type is as follows:
wherein W is u Represents the degree of interest, M represents the number of species, N ui Represent the degree of interest of type i, N u Representing a sum of the attention degrees corresponding to each type; the second calculation module is used for calculating a customized recommended content matrix according to the interest extensive degree of the user and a preset standard recommended content matrix, wherein the standard recommended content matrix and the same position element in the customized recommended content matrix correspond to the same standard recommended content, and the value of each element in the customized recommended content matrix is used for representing the recommended priority of the standard recommended content;
And the selecting module is used for selecting recommended content from a plurality of standard recommended content based on the recommended priority.
4. The content recommendation device of claim 3 wherein said second computing module comprises:
the decomposition unit is used for reversely decomposing the standard recommended content matrix to obtain a correlation matrix and a similarity matrix;
the second calculation unit is used for calculating the product of the interest extensive degree and the similarity matrix to obtain a customized similarity matrix;
and the generation unit is used for generating the customized recommended content matrix according to the customized similarity matrix and the correlation matrix.
5. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any of claims 1-2 when executing a program stored on a memory.
6. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any of claims 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010144164.7A CN111291217B (en) | 2020-03-04 | 2020-03-04 | Content recommendation method, device, electronic equipment and computer readable medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010144164.7A CN111291217B (en) | 2020-03-04 | 2020-03-04 | Content recommendation method, device, electronic equipment and computer readable medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111291217A CN111291217A (en) | 2020-06-16 |
| CN111291217B true CN111291217B (en) | 2024-02-02 |
Family
ID=71024723
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010144164.7A Active CN111291217B (en) | 2020-03-04 | 2020-03-04 | Content recommendation method, device, electronic equipment and computer readable medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111291217B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112395496A (en) * | 2020-10-22 | 2021-02-23 | 上海众源网络有限公司 | Information recommendation method and device, electronic equipment and storage medium |
| CN112351345A (en) * | 2020-11-04 | 2021-02-09 | 深圳Tcl新技术有限公司 | Control method and device of recommended content, smart television and storage medium |
| CN116029798B (en) * | 2023-03-22 | 2023-07-07 | 北京新发地农产品网络配送中心有限责任公司 | User demand recommendation method, system, electronic equipment and readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102866997A (en) * | 2011-07-05 | 2013-01-09 | 腾讯科技（深圳）有限公司 | Method and device for processing user data |
| CN109697629A (en) * | 2018-11-15 | 2019-04-30 | 平安科技（深圳）有限公司 | Product data method for pushing and device, storage medium, computer equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10191949B2 (en) * | 2015-06-18 | 2019-01-29 | Nbcuniversal Media, Llc | Recommendation system using a transformed similarity matrix |
-
2020
- 2020-03-04 CN CN202010144164.7A patent/CN111291217B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102866997A (en) * | 2011-07-05 | 2013-01-09 | 腾讯科技（深圳）有限公司 | Method and device for processing user data |
| CN109697629A (en) * | 2018-11-15 | 2019-04-30 | 平安科技（深圳）有限公司 | Product data method for pushing and device, storage medium, computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111291217A (en) | 2020-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110321422B (en) | Method for training model on line, pushing method, device and equipment | |
| CN110929052B (en) | Multimedia resource recommendation method and device, electronic equipment and storage medium | |
| CN111291217B (en) | Content recommendation method, device, electronic equipment and computer readable medium | |
| CN109086439B (en) | Information recommendation method and device | |
| WO2017181612A1 (en) | Personalized video recommendation method and device | |
| US20170188102A1 (en) | Method and electronic device for video content recommendation | |
| CN110807207B (en) | Data processing method and device, electronic equipment and storage medium | |
| CN110717099B (en) | Method and terminal for recommending film | |
| CN109903103B (en) | Method and device for recommending articles | |
| CN111010592B (en) | Video recommendation method and device, electronic equipment and storage medium | |
| CN110489574B (en) | Multimedia information recommendation method and device and related equipment | |
| CN112507163B (en) | Duration prediction model training method, recommendation method, device, equipment and medium | |
| US20220261591A1 (en) | Data processing method and apparatus | |
| CN102165441A (en) | Method, system, and apparatus for ranking media sharing channels | |
| US20190278819A1 (en) | Systems and methods of providing recommendations of content items | |
| CN107346333B (en) | Online social network friend recommendation method and system based on link prediction | |
| CN115455280A (en) | Recommendation list determining method and server | |
| CN110085292B (en) | Medicine recommendation method and device and computer-readable storage medium | |
| CN112989118A (en) | Video recall method and device | |
| CN115730217A (en) | Model training method, material recalling method and device | |
| CN111984867B (en) | Network resource determining method and device | |
| CN111552883B (en) | Content recommendation method and computer-readable storage medium | |
| CN110275986B (en) | Video recommendation method based on collaborative filtering, server and computer storage medium | |
| CN110309361B (en) | Video scoring determination method, recommendation method and device and electronic equipment | |
| Gan | TAFFY: incorporating tag information into a diffusion process for personalized recommendations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |