CN117235366B - Collaborative recommendation method and system based on content relevance - Google Patents
Collaborative recommendation method and system based on content relevance Download PDFInfo
- Publication number
- CN117235366B CN117235366B CN202311204350.5A CN202311204350A CN117235366B CN 117235366 B CN117235366 B CN 117235366B CN 202311204350 A CN202311204350 A CN 202311204350A CN 117235366 B CN117235366 B CN 117235366B
- Authority
- CN
- China
- Prior art keywords
- content
- user
- click
- scoring matrix
- obtaining
- 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 64
- 239000011159 matrix material Substances 0.000 claims abstract description 67
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 17
- 238000000342 Monte Carlo simulation Methods 0.000 claims abstract description 14
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 14
- 239000006185 dispersion Substances 0.000 claims abstract description 14
- 230000006399 behavior Effects 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 claims description 3
- 238000011946 reduction process Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 abstract description 6
- 238000001914 filtration Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 235000002198 Annona diversifolia Nutrition 0.000 description 1
- 241000282842 Lama glama Species 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention discloses a collaborative recommendation method and a collaborative recommendation system based on content relevance, which calculate similarity sim ij between video contents; performing dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user; calculating to obtain an acceptance set A i of the user u on the related content set based on the content preference weight vector set COE i; simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click extension set CKI of the user; constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click extension set CKI; and obtaining a predictive scoring matrix of the content scoring matrix by using a singular value decomposition algorithm, and obtaining a content recommendation list based on the predictive scoring matrix. The collaborative recommendation method based on the content relevance solves the problems that an existing recommendation system cannot reflect the content relevance of recommendation, user experience is poor, and page conversion rate is low.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a collaborative recommendation method, a collaborative recommendation system, electronic equipment and a storage medium based on content relevance.
Background
Along with the rapid development of the mobile internet, the scale of internet users is approaching to the growing vertex, and when massive information fills our lives, it is very important that valuable information can be obtained efficiently and rapidly;
The existing recommendation system is used as an efficient information filtering tool, so that a user can accurately and efficiently acquire information, collaborative filtering is the most widely applied recommendation algorithm, and the recommendation system plays an extremely important role in the recommendation field. Collaborative filtering simply uses the preference of a certain interest group to recommend information of interest to a user, and individuals give considerable responses to the information through a collaborative mechanism and record the responses so as to achieve the purpose of filtering, thereby helping others to screen the information. The collaborative filtering method has the advantages that domain knowledge is not needed, the algorithm is easy to realize in a distributed mode, and massive data sets can be processed. However, the collaborative filtering method has the defects of sparse data and incapability of reflecting the correlation of recommended contents.
What is needed is a collaborative recommendation method that reflects the relevance of recommended content, and that can improve user experience and increase page conversion.
Disclosure of Invention
The embodiment of the invention aims to provide a collaborative recommendation method, a collaborative recommendation system, electronic equipment and a storage medium based on content relevance, which are used for solving the problems that the existing recommendation system cannot reflect the relevance of recommended content, the user experience is poor and the page conversion rate is low.
In order to achieve the above object, an embodiment of the present invention provides a collaborative recommendation method based on content relevance, where the method specifically includes:
S101, acquiring a plurality of video contents, and calculating similarity sim ij between the video contents;
s102, obtaining a content click set CK i based on click behavior data of a user u in a preset time, and performing dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user;
s103, acquiring a related content set CR i related to all videos in the content click set CK i, and calculating to obtain a receptivity set A i of a user u on the related content set based on the content preference weight vector set COE i;
s104, simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click expansion set CK I of the user;
S105, repeating the steps S102-S104 to obtain a content click set CK and a content click expansion set CK I of all users, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click expansion set CK I;
S106, obtaining a predictive scoring matrix of the content scoring matrix by using a singular value decomposition algorithm, and obtaining a content recommendation list based on the predictive scoring matrix.
Based on the technical scheme, the invention can also be improved as follows:
Further, the obtaining a plurality of video contents and calculating the similarity sim ij between the video contents includes:
Extracting text data of each video content by an ASR technology, and obtaining text description text i corresponding to each video content based on the text data and the video content attribute;
And calculating the similarity sim ij between the text i and the text j, and obtaining the similarity sim ij between the video content i and the video content j based on the similarity sim ij.
Further, the obtaining a content click set CK i based on click behavior data of the user u in a preset time, performing dispersion normalization on the content click set CK i, and obtaining a content preference weight vector set COE i of the user, includes:
Calculating a content preference weight vector of a user u through a formula 1;
In the method, in the process of the invention, For content preference weight vector,CK i is the content click set.
Further, the simulating the clicking condition of the user on the related content set CR i by using the monte carlo simulation method to obtain the content clicking expansion set CK I of the user includes:
Constructing a probability set P i of clicking of the related content set CR i by the user u through the receptivity set A i of the user u;
And calculating to obtain a content click expansion set CK I of the user based on the probability set P i.
Further, the repeating steps S102-S104 obtain a content click set CK and a content click expansion set CK I of all users, and construct a content scoring matrix R n×m of the current user based on the content click set CK and the content click expansion set CK I, including:
Decomposing the content scoring matrix R n×m;
And performing dimension reduction on the content scoring matrix R n×m subjected to decomposition.
Further, repeating the steps S102-S104 to obtain a content click set CK and a content click extension set CKI of all users, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click extension set CK I, further including:
after the dimension reduction process, calculating a recommendation score for recommending the video content i to the user u through a formula 2;
In the method, in the process of the invention, To recommend video content i to user u, p u is the feature vector of user u,/>, a recommendation scoreIs a feature vector of video content i.
Further, the collaborative recommendation method based on the content relevance further comprises the following steps:
And sequencing the recommendation scores to obtain a content recommendation list, and generating a recommendation page based on the content recommendation list.
A collaborative recommendation system based on content relevance, comprising:
The acquisition module is used for acquiring a plurality of video contents;
The similarity calculation module is used for calculating similarity sim ij between the video contents;
The content click set determining module is used for obtaining a content click set CK i based on click behavior data of a user u in a preset time;
The content preference weight vector set determining module is used for carrying out dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user;
The related content set determining module is used for obtaining a related content set CR i related to all videos in the content click set CK i;
The acceptance degree set determining module is used for calculating an acceptance degree set A i of the user u on the related content set based on the content preference weight vector set COE i;
the content click expansion set determining module is used for simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click expansion set CK I of the user;
The content scoring matrix determining module is used for obtaining a content clicking set CK and a content clicking expansion set CK I of all users and constructing a content scoring matrix R n×m of the current user based on the content clicking set CK and the content clicking expansion set CK I;
and the content recommendation list determining module is used for obtaining a prediction scoring matrix of the content scoring matrix by using a singular value decomposition algorithm and obtaining a content recommendation list based on the prediction scoring matrix.
An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method when the computer program is executed.
A non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method.
The embodiment of the invention has the following advantages:
According to the collaborative recommendation method based on the content relevance, a plurality of video contents are obtained, and similarity sim ij between the video contents is calculated; obtaining a content click set CK i based on click behavior data of a user u in a preset time, and performing dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user; acquiring a related content set CR i related to all videos in the content click set CK i, and calculating to obtain a user u acceptance set A i of the related content set based on the content preference weight vector set COE i; simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click expansion set CK I of the user; repeating the steps S102-S104 to obtain a content click set CK and a content click expansion set CKI of all users, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click expansion set CK I; the predictive scoring matrix of the content scoring matrix is obtained by using a singular value decomposition algorithm, and a content recommendation list is obtained based on the predictive scoring matrix, so that the problems that a recommendation system cannot reflect the relevance of recommended content, the user experience is poor and the page conversion rate is low in the prior art are solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those skilled in the art from this disclosure that the drawings described below are merely exemplary and that other embodiments may be derived from the drawings provided without undue effort.
The structures, proportions, sizes, etc. shown in the present specification are shown only for the purposes of illustration and description, and are not intended to limit the scope of the invention, which is defined by the claims, so that any structural modifications, changes in proportions, or adjustments of sizes, which do not affect the efficacy or the achievement of the present invention, should fall within the scope of the invention.
FIG. 1 is a flow chart of a collaborative recommendation method based on content relevance according to the present invention;
FIG. 2 is a block diagram of a collaborative recommendation system based on content relevance according to the present invention;
fig. 3 is a schematic diagram of an entity structure of an electronic device according to the present invention.
Wherein the reference numerals are as follows:
The system comprises an acquisition module 10, a similarity calculation module 20, a content click set determination module 30, a content preference weight vector set determination module 40, a related content set determination module 50, an acceptance set determination module 60, a content click expansion set determination module 70, a content scoring matrix determination module 80, a content recommendation list determination module 90, an electronic device 100, a processor 1001, a memory 1002 and a bus 1003.
Detailed Description
Other advantages and advantages of the present invention will become apparent to those skilled in the art from the following detailed description, which, by way of illustration, is to be read in connection with certain specific embodiments, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Examples
Fig. 1 is a flowchart of an embodiment of a collaborative recommendation method based on content relevance according to the present invention, and as shown in fig. 1, the collaborative recommendation method based on content relevance according to the embodiment of the present invention includes the following steps:
s101, acquiring a plurality of video contents, and calculating the similarity between the video contents;
Specifically, extracting text data of each video content by an ASR technology, and obtaining text description text i corresponding to each video content based on the text data and the video content attribute; video content attributes include domain, author, organization, etc.;
table 1 is a content text description table; as shown in table 1;
TABLE 1
The similarity sim ij between the text description text i and the text description text j is calculated through the large language model LLaMA, and the larger the similarity sim ij,simij∈[0,1),simij between the video content i and the video content j is obtained based on the similarity sim ij, the higher the correlation between the video content i and the video content j is, and sim ij =0 indicates that there is no correlation between the video content i and the video content j.
Table 2 is a table of content correlation:
TABLE 2
S102, obtaining a content click set based on click behavior data of a user u in a preset time, and performing dispersion standardization on the content click set to obtain a content preference weight vector set of the user;
Specifically, a content preference weight vector of a user u is calculated through a formula 1;
In the method, in the process of the invention, For content preference weight vector,CK i is the content click set.
Preferably, the preset time is about one month, and the content clicks are collectedDefining a content set C i as a user click content set;
table 3 is a user click behavior table;
TABLE 3 Table 3
Table 4 is the preference weight of the user for the content topic;
TABLE 4 Table 4
S103, acquiring a related content set which has relevance with all videos in the content click set, and calculating based on the content preference weight vector set to obtain an acceptance set of the user u on the related content set;
Specifically, the video content in CR i is video content that has not been clicked in the last month of the user and if j e C i、u∈CRi, sim ju e (0, 1) exists;
namely, CR i is a content set which is not clicked in the last month of the user and has relevance to the content clicked by the user, and for j epsilon C i of the user u, a relevant content set of u epsilon CR i is clicked, and a relevance set SIM u＝{…simju…},j∈Ci of u exists; the acceptance of user u for content u e CR i is:
j，l∈Ci,simju,simlu∈SIMu；
The acceptance set of the user u on the related content set CR i is obtained according to the above u，k∈CRi；
Table 5 is a table of related content receptivity;
TABLE 5
S104, simulating the clicking condition of the user on the related content set by using a Monte Carlo simulation method to obtain a content clicking expansion set CK I of the user.
Specifically, there are multiple possible outcomes for each trial, but only one will occur for each outcome;
each test is independent, and each test result is not influenced by other test results.
Constructing a probability set P i of clicking of the related content set CR i by the user u through the receptivity set A i of the user u; as can be seen, SUM (P i) =1;
From the probability set P i, it is known that each test result has a respective occurrence probability, and the sum of occurrence probabilities of all the results is 1.
From the above, the user u presents a plurality of distributions to the clicks of the content in the related content set CR i, and the probability of each click result is P i; thus, a Monte Carlo simulation method can be used for experimental simulation. Performing n times of experimental simulation to obtain a content click expansion set of the user u as
Table 6 is a simulated user click behavior data table;
TABLE 6
| Number of tests n | User ud | Time of completion of test | Content id |
| 1 | f0d2... | 2023/7/25 12:44 | 6dbd40d1 |
| 2 | f0d2... | 2023/7/25 12:44 | 5bbdffda |
| 3 | f0d2... | 2023/7/25 12:44 | cfc7daa8 |
| 4 | f0d2... | 2023/7/25 12:44 | f5ff433d |
| 5 | f0d2... | 2023/7/25 12:44 | 4dee9bbe |
| 6 | f0d2... | 2023/7/25 12:44 | 10fa82ff |
| 7 | f0d2... | 2023/7/25 12:44 | 7e12e8a0 |
| 8 | f0d2... | 2023/7/25 12:45 | c1a16998 |
| 9 | f0d2... | 2023/7/25 12:45 | 71998c29 |
| 10 | f0d2... | 2023/7/25 12:45 | d77f0009 |
| 11 | f0d2... | 2023/7/25 12:45 | 34d3d3f2 |
| 12 | f0d2... | 2023/7/25 12:45 | 6732d340 |
| ... | f0d2... | ... | ... |
| 6732 | f0d2... | 2023/7/25 12:48 | 4bcabcea |
| 6733 | f0d2... | 2023/7/25 12:48 | 6c8a6b12 |
| 6734 | f0d2... | 2023/7/25 12:48 | 5572426d |
| 6735 | f0d2... | 2023/7/25 12:48 | 6c5daf3a |
| 6736 | f0d2... | 2023/7/25 12:48 | a8f6f7a8 |
| 6737 | f0d2... | 2023/7/25 12:48 | 5210c843 |
| 6738 | f0d2... | 2023/7/25 12:48 | 4d0d9bc6 |
| 6739 | f0d2... | 2023/7/25 12:48 | 76edc84e |
| 6740 | f0d2... | 2023/7/25 12:48 | bee99628 |
| 6741 | f0d2... | 2023/7/25 12:48 | 50aaf1cb |
S105, repeating the steps S102-S104 to obtain a content click set CK and a content click expansion set CK I of all users, and constructing a content scoring matrix of the current user based on the content click set CK and the content click expansion set CK I;
specifically, the content scoring matrix R n×m is decomposed;
r=uΣv T, where U is an orthogonal matrix of n, V is an orthogonal matrix of m, Σn is a matrix of n.
And performing dimension reduction on the content scoring matrix R n×m subjected to decomposition treatment, and reserving a part of larger singular values, wherein the diagonal elements are assigned with 0.
The method comprises the following steps of: The dimension reduction is to map the user and the content to a k dimension space in a physical sense, and the recommendation score can be calculated by the dot product of the user vector and the content vector after the dimension reduction,
Calculating a recommendation score for recommending the video content i to the user u through a formula 2;
In the method, in the process of the invention, To recommend video content i to user u, p u is the feature vector of user u,/>, a recommendation scoreIs a feature vector of video content i.
Loss function according to singular value decomposition:
The problem of mapping users and contents into a k-dimensional space is converted into a machine learning problem, and a common solving method comprises a random gradient descent method and an alternate least square method;
Table 7 is a user-content scoring matrix;
TABLE 7
S106, obtaining a predictive scoring matrix of the content scoring matrix by using a singular value decomposition algorithm, and obtaining a content recommendation list based on the predictive scoring matrix;
Specifically, sorting the predictive scores of the unscored parts of the target users, and selecting the content with the highest predictive score to form a recommendation list;
Table 8 is a user recommendation list;
TABLE 8
And sequencing the recommendation scores to obtain a content recommendation list, and generating a recommendation page based on the content recommendation list.
According to the collaborative recommendation method based on the content relevance, a plurality of video contents are obtained, and similarity sim ij between the video contents is calculated; obtaining a content click set CK i based on click behavior data of a user u in a preset time, and performing dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user; acquiring a related content set CR i related to all videos in the content click set CK i, and calculating to obtain a user u acceptance set A i of the related content set based on the content preference weight vector set COE i; simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click expansion set CK I of the user; repeating the steps S102-S104 to obtain a content click set CK and a content click expansion set CK I of all users, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click expansion set CK I; and obtaining a predictive scoring matrix of the content scoring matrix by using a singular value decomposition algorithm, and obtaining a content recommendation list based on the predictive scoring matrix. The problem that the existing recommendation system cannot reflect the correlation of recommended content, and is poor in user experience and low in page conversion rate is solved.
FIG. 2 is a flowchart of a collaborative recommendation system based on content relevance according to an embodiment of the present invention; as shown in fig. 2, the collaborative recommendation system based on content relevance provided by the embodiment of the invention includes the following steps:
An acquisition module 10, configured to acquire a plurality of video contents;
A similarity calculating module 20, configured to calculate a similarity sim ij between the video contents;
The content click set determining module 30 is configured to obtain a content click set CK i based on click behavior data of the user u in a preset time;
The content preference weight vector set determining module 40 is configured to perform dispersion normalization on the content click set CK i to obtain a content preference weight vector set COE i of the user;
a related content set determining module 50, configured to obtain a related content set CR i related to all videos in the content click set CK i;
the acceptance set determining module 60 is configured to calculate, based on the content preference weight vector set COE i, an acceptance set a i of the user u for the related content set;
The content click expansion set determining module 70 is configured to simulate a click condition of a user on the related content set CR i by using a monte carlo simulation method, so as to obtain a content click expansion set CK I of the user;
The content scoring matrix determining module 80 is configured to obtain a content click set CK and a content click extension set CK I of all users, and construct a content scoring matrix R n×m of the current user based on the content click set CK and the content click extension set CK I;
The content recommendation list determining module 90 is configured to obtain a prediction scoring matrix of the content scoring matrix using a singular value decomposition algorithm, and obtain a content recommendation list based on the prediction scoring matrix.
According to the collaborative recommendation system based on the content relevance, a plurality of video contents are acquired through an acquisition module 10; calculating a similarity sim ij between the video contents by a similarity calculation module 20; obtaining a content click set CK i based on click behavior data of a user u in a preset time through a content click set determining module 30; performing dispersion normalization on the content click set CK i through a content preference weight vector set determining module 40 to obtain a content preference weight vector set COE i of the user; acquiring a related content set CR i related to all videos in the content click set CK i through a related content set determination module 50; calculating, by an acceptance set determining module 60, an acceptance set a i of the related content set by the user u based on the content preference weight vector set COE i; simulating the click condition of a user on a related content set CR i by using a Monte Carlo simulation method through a content click expansion set determining module 70 to obtain a content click expansion set CK I of the user; obtaining a content click set CK and a content click expansion set CK I of all users through a content scoring matrix determining module 80, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click expansion set CK I; a predictive scoring matrix for the content scoring matrix is obtained by the content recommendation list determination module 90 using a singular value decomposition algorithm, and a content recommendation list is obtained based on the predictive scoring matrix. The collaborative recommendation method based on the content relevance solves the problems that an existing recommendation system cannot reflect the content relevance of recommendation, user experience is poor, and page conversion rate is low.
Fig. 3 is a schematic diagram of an entity structure of an electronic device according to an embodiment of the present invention, as shown in fig. 3, an electronic device 100 includes: a processor 1001, a memory 1002, and a bus 1003;
Wherein, the processor 1001 and the memory 1002 complete communication with each other through the bus 1003;
The processor 1001 is configured to call program instructions in the memory 1002 to perform the methods provided by the above method embodiments, for example, including: acquiring a plurality of video contents, and calculating similarity sim ij between the video contents; obtaining a content click set CK i based on click behavior data of a user u in a preset time, and performing dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user; acquiring a related content set CR i related to all videos in the content click set CK i, and calculating to obtain a user u acceptance set A i of the related content set based on the content preference weight vector set COE i; simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click extension set CKI of the user; obtaining a content click set CK and a content click extension set CKI of all users, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click extension set CKI; and obtaining a predictive scoring matrix of the content scoring matrix by using a singular value decomposition algorithm, and obtaining a content recommendation list based on the predictive scoring matrix.
The present embodiment provides a non-transitory computer readable storage medium storing computer instructions that cause a computer to perform the methods provided by the above-described method embodiments, for example, including: acquiring a plurality of video contents, and calculating similarity sim ij between the video contents; obtaining a content click set CK i based on click behavior data of a user u in a preset time, and performing dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user; acquiring a related content set CR i related to all videos in the content click set CK i, and calculating to obtain a user u acceptance set A i of the related content set based on the content preference weight vector set COE i; simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click extension set CKI of the user; obtaining a content click set CK and a content click extension set CKI of all users, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click extension set CKI; and obtaining a predictive scoring matrix of the content scoring matrix by using a singular value decomposition algorithm, and obtaining a content recommendation list based on the predictive scoring matrix.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware associated with program instructions, where the foregoing program may be stored in a computer readable storage medium, and when executed, the program performs steps including the above method embodiments; and the aforementioned storage medium includes: various storage media such as ROM, RAM, magnetic or optical disks may store program code.
The apparatus embodiments described above are merely illustrative, wherein elements illustrated as separate elements may or may not be physically separate, and elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on such understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the embodiments or the methods of some parts of the embodiments.
While the invention has been described in detail in the foregoing general description and specific examples, it will be apparent to those skilled in the art that modifications and improvements can be made thereto. Accordingly, such modifications or improvements may be made without departing from the spirit of the invention and are intended to be within the scope of the invention as claimed.
Claims (7)
1. The collaborative recommendation method based on the content relevance is characterized by comprising the following steps:
S101, acquiring a plurality of video contents, and calculating similarity sim ij between the video contents;
s102, obtaining a content click set CK i based on click behavior data of a user u in a preset time, and performing dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user;
S103, acquiring a related content set CR i related to all videos in the content click set CK i, and if j epsilon C i、u∈CRi, sim ju epsilon (0, 1) exists;
And based on the content preference weight vector set COE i, calculating to obtain a receptivity set a i of the user u to the related content set, u，k∈CRi；
S104, simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click extension set CKI of the user;
Constructing a probability set P i of clicking of the user u on the related content set CR through the receptivity set A i of the user u;
Calculating to obtain a content click extension set CKI of the user based on the probability set P i;
S105, repeating the steps S102-S104 to obtain a content click set CK and a content click extension set CKI of all users, and constructing a content scoring matrix R n×m of the current user based on the content click set CK and the content click extension set CKI;
Decomposing the content scoring matrix R n×m;
performing dimension reduction on the content scoring matrix R n×m subjected to decomposition treatment;
after the dimension reduction process, calculating a recommendation score for recommending the video content i to the user u through a formula 2;
In the method, in the process of the invention, To recommend video content i to user u, p u is the feature vector of user u,/>, a recommendation scoreIs a feature vector of video content i;
S106, obtaining a predictive scoring matrix of the content scoring matrix by using a singular value decomposition algorithm, and obtaining a content recommendation list based on the predictive scoring matrix.
2. The collaborative recommendation method according to claim 1, wherein the obtaining a plurality of video contents and calculating a similarity sim ij between the video contents comprises:
Extracting text data of each video content by an ASR technology, and obtaining text description text i corresponding to each video content based on the text data and the video content attribute;
And calculating the similarity sim ij between the text i and the text j, and obtaining the similarity sim ij between the video content i and the video content j based on the similarity sim ij.
3. The collaborative recommendation method based on content relevance according to claim 1, wherein the obtaining a content click set CK i based on click behavior data of a user u in a preset time, performing dispersion normalization on the content click set CK i, and obtaining a content preference weight vector set COE i of the user, includes:
Calculating a content preference weight vector of a user u through a formula 1;
In the method, in the process of the invention, For content preference weight vector,CK i is the content click set and i is the video content.
4. The collaborative recommendation method based on content relevance according to claim 3, further comprising:
And sequencing the recommendation scores to obtain a content recommendation list, and generating a recommendation page based on the content recommendation list.
5. A collaborative recommendation system based on content relevance, comprising:
The acquisition module is used for acquiring a plurality of video contents;
The similarity calculation module is used for calculating similarity sim ij between the video contents;
The content click set determining module is used for obtaining a content click set CK i based on click behavior data of a user u in a preset time;
The content preference weight vector set determining module is used for carrying out dispersion standardization on the content click set CK i to obtain a content preference weight vector set COE i of the user;
the related content set determining module is used for obtaining a related content set CR i related to all videos in the content click set CK i, and if j epsilon C i、u∈CRi, sim ju epsilon (0, 1) exists;
an acceptance degree set determining module, configured to calculate an acceptance degree set a i of the user u for the related content set based on the content preference weight vector set COE i, u，k∈CRi；
The content click extension set determining module is used for simulating the click condition of a user on the related content set CR i by using a Monte Carlo simulation method to obtain a content click extension set CKI of the user;
Constructing a probability set P i of clicking of the related content set CR i by the user u through the receptivity set A i of the user u;
Calculating to obtain a content click extension set CKI of the user based on the probability set P i;
The content scoring matrix determining module is used for obtaining a content clicking set CK and a content clicking extension set CKI of all users and constructing a content scoring matrix R n×m of the current user based on the content clicking set CK and the content clicking extension set CKI;
Decomposing the content scoring matrix R n×m;
performing dimension reduction on the content scoring matrix R n×m subjected to decomposition treatment;
after the dimension reduction process, calculating a recommendation score for recommending the video content i to the user u through a formula 2;
In the method, in the process of the invention, To recommend video content i to user u, p u is the feature vector of user u,/>, a recommendation scoreIs a feature vector of video content i;
and the content recommendation list determining module is used for obtaining a prediction scoring matrix of the content scoring matrix by using a singular value decomposition algorithm and obtaining a content recommendation list based on the prediction scoring matrix.
6. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1 to 4 when the computer program is executed.
7. A non-transitory computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311204350.5A CN117235366B (en) | 2023-09-19 | 2023-09-19 | Collaborative recommendation method and system based on content relevance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311204350.5A CN117235366B (en) | 2023-09-19 | 2023-09-19 | Collaborative recommendation method and system based on content relevance |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117235366A CN117235366A (en) | 2023-12-15 |
| CN117235366B true CN117235366B (en) | 2024-06-18 |
Family
ID=89092552
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311204350.5A Active CN117235366B (en) | 2023-09-19 | 2023-09-19 | Collaborative recommendation method and system based on content relevance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117235366B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107515909A (en) * | 2017-08-11 | 2017-12-26 | 深圳市耐飞科技有限公司 | A kind of video recommendation method and system |
| CN108804683A (en) * | 2018-06-13 | 2018-11-13 | 重庆理工大学 | Associate(d) matrix decomposes and the film of collaborative filtering recommends method |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104063481B (en) * | 2014-07-02 | 2017-11-14 | 山东大学 | A kind of film personalized recommendation method based on the real-time interest vector of user |
| KR20160064447A (en) * | 2014-11-28 | 2016-06-08 | 이종찬 | A recommendation method for new users by using preference prediction based on collaborative filtering algorithm |
| CN109947983A (en) * | 2017-09-19 | 2019-06-28 | Tcl集团股份有限公司 | Video recommendation method, system, terminal and computer readable storage medium |
| CN114912011A (en) * | 2021-12-27 | 2022-08-16 | 天翼数字生活科技有限公司 | Video recommendation method based on content extraction and scoring prediction |
-
2023
- 2023-09-19 CN CN202311204350.5A patent/CN117235366B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107515909A (en) * | 2017-08-11 | 2017-12-26 | 深圳市耐飞科技有限公司 | A kind of video recommendation method and system |
| CN108804683A (en) * | 2018-06-13 | 2018-11-13 | 重庆理工大学 | Associate(d) matrix decomposes and the film of collaborative filtering recommends method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117235366A (en) | 2023-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kosinski et al. | Mining big data to extract patterns and predict real-life outcomes. | |
| Cox et al. | Working with missing data in higher education research: A primer and real-world example | |
| Hagemann | Cluster-robust bootstrap inference in quantile regression models | |
| Yang-Wallentin et al. | Confirmatory factor analysis of ordinal variables with misspecified models | |
| US11636314B2 (en) | Training neural networks using a clustering loss | |
| US11842257B2 (en) | System and method for predicting and scoring a data model | |
| WO2022048363A1 (en) | Website classification method and apparatus, computer device, and storage medium | |
| US20130204835A1 (en) | Method of extracting named entity | |
| JP2017199355A (en) | Recommendation generation | |
| US8086555B2 (en) | Collaborative filtering systems and methods | |
| Karavias et al. | Local power of panel unit root tests allowing for structural breaks | |
| Chen et al. | A model averaging approach for the ordered probit and nested logit models with applications | |
| CN117235366B (en) | Collaborative recommendation method and system based on content relevance | |
| CN117573985B (en) | Information pushing method and system applied to intelligent online education system | |
| CN112836754A (en) | Image description model generalization capability evaluation method | |
| CN111552810B (en) | Entity extraction and classification method, entity extraction and classification device, computer equipment and storage medium | |
| CN117649464A (en) | Weak supervision image synthesis method, system, equipment and storage medium | |
| US20230214451A1 (en) | System and method for finding data enrichments for datasets | |
| CN106294584A (en) | The training method of order models and device | |
| CN113505838B (en) | Image clustering method and device, electronic equipment and storage medium | |
| CN111581382B (en) | Method and system for predicting hot questions in question-answering community | |
| CN116091113A (en) | Marketing model data processing method, system and computer readable storage medium | |
| CN114491296A (en) | Proposal affiliate recommendation method, system, computer device and readable storage medium | |
| JP6509391B1 (en) | Computer system | |
| CN113282738B (en) | Text selection method and device |
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 |