CN111143699B - Recommendation system based on similarity and confidence coefficient clustering - Google Patents

Recommendation system based on similarity and confidence coefficient clustering Download PDF

Info

Publication number
CN111143699B
CN111143699B CN202010006918.2A CN202010006918A CN111143699B CN 111143699 B CN111143699 B CN 111143699B CN 202010006918 A CN202010006918 A CN 202010006918A CN 111143699 B CN111143699 B CN 111143699B
Authority
CN
China
Prior art keywords
user
users
similarity
neighbor
cluster
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
Application number
CN202010006918.2A
Other languages
Chinese (zh)
Other versions
CN111143699A (en
Inventor
王佳伟
苏湛
艾均
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Shanghai for Science and Technology
Original Assignee
University of Shanghai for Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Shanghai for Science and Technology filed Critical University of Shanghai for Science and Technology
Priority to CN202010006918.2A priority Critical patent/CN111143699B/en
Publication of CN111143699A publication Critical patent/CN111143699A/en
Application granted granted Critical
Publication of CN111143699B publication Critical patent/CN111143699B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • 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

Abstract

The invention provides a recommendation system based on similarity and confidence coefficient clustering, which takes the similarity and confidence coefficient between users in the recommendation system as consideration factors, and screens similar users by k-means clustering as a selection range, so that the diversity of articles recommended by the recommendation system is improved, the accuracy of recommendation is improved, and besides, the number of neighbors required by optimal recommendation is reduced.

Description

Recommendation system based on similarity and confidence coefficient clustering
Technical Field
The invention relates to the technical field of recommendation systems, in particular to a recommendation system based on similarity and confidence clustering.
Background
Recommendation systems are being focused and studied as hot spot problems by many research areas. It may find a set of users with similar interests based on the interests of the users and then recommend items of interest to the users based on the similarities between the users. The personalized recommendation technology is used for recommending movies, commodities and the like suitable for the user, so that the user can quickly obtain item information conforming to liking and selecting trends, and the recommendation system can be trusted by the user to obtain more liking and use, and the system recommendation saves system resources consumed by the user in a large amount of browsing processes, thereby saving operation cost such as bandwidth and the like and better serving the user.
As the recommendation technology which is most successfully applied in the recommendation system, the conventional collaborative filtering recommendation algorithm has two problems. The method only depends on the similarity between users to select the neighbors of the target user, so that the recommendation accuracy is low, and the defect of poor diversity of recommended commodities exists during recommendation.
Disclosure of Invention
The invention aims to provide a recommendation system based on similarity and confidence clustering, which is used for selecting more proper neighbors according to two indexes and needs fewer neighbors to realize more accurate similarity calculation and recommendation service.
In order to achieve the above purpose, the present invention provides a recommendation system based on similarity and confidence clustering, comprising the following steps:
step 1: establishing a user scoring library in a current recommendation system;
step 2: calculating the similarity of a target user Ua and other users based on the scores of the users on different articles, and generating a set Si;
step 3: based on the relation between users in Si, taking the jointly scored articles between the target user Ua and the users as confidence coefficients to generate a set Ri;
step 4: placing the users in a two-dimensional coordinate system based on two indexes of similarity between the users in Si and confidence coefficients between the users in Ri;
step 5: dividing users into two types according to a K-means clustering method, and taking the type with higher similarity and confidence coefficient in the clustered two types as a similar neighbor set;
step 6: dividing neighbors in the similar neighbor set into high-reliability neighbors and low-reliability neighbors and giving different weights to the two neighbors;
step 7: calculating the predictive score of the user on the item Ia according to the score of the similar neighbor of the user on the item Ia and the weight of the similar neighbor; and recommending the result to the user Ua according to the score.
In step 2, the following steps are included:
step 2.1: the user scoring data is normalized with the following formula:
wherein r is i max And r i min Representing the highest score and the lowest score in the user i score record, if the highest score is equal to the lowest score, the normalized value can be assigned to be 0, eiαepsilon < -1,1 []
Step 2.2: the target user Ua searches for and/or scores item Ia;
step 2.3: for any two users i and j, the correlation calculation formula is as follows
Through the formula, integrating correlation data between the target user Ua and other users to generate a set Si;
wherein k is α Representing the degree, k, of an article i Respectively representing the degrees of users; s is S ij The influence of user i and user j is expressed and can be regarded as the similarity of user i and user j, but the similarity of the users is directional, namely the influence between two articles is different, and the user i scores the article alpha excessively, then a iα 1, otherwise 0.
In step 3, the following steps are included:
step 3.1: for any two users i and j, the confidence coefficient calculation formula is as follows:
through the formula, integrating confidence coefficients between the target user Ua and other users to generate a set Ri;
wherein n is ij The reason for introducing this index for the number of items commonly scored by users i and j is as follows, assuming that users i and j commonly rated 100 items and users i and k commonly rated 10 items, then the similarity S between users i and j ij Will be compared with the similarity S between i and k ik And more credibility is realized.
In step 4, the following steps are included:
step 4.1: with similarity S between users ij Confidence coefficient n ij As an evaluation index, a two-dimensional coordinate system is established, and a user is taken as a node.
In step 5, the following steps are included:
step 5.1: sample set X i ＝{(S i1 ，f i1 )，(S i2 ，f i2 )，(S i3 ，f i3 )，...，(S ij ，f ij ) Finding out the maximum similarity between user i and other usersMinimum similarity->And maximum confidence coefficient->Minimum confidence coefficient->Handle->And->Initial centroid μ as cluster 1 Sum mu 2 ；
Step 5.2: calculating the distance between each node and centroid muThe formula is as follows:
wherein S is μ Similarity of centroid μ, f μ Confidence coefficient for centroid μ, from node j toCentroid mu 1 Sum mu 2 Distance of (2)And->The size of (2) divides the points in the coordinate system into two clusters.
Step 5.3: according to the cluster C after division 1 And C 2 In the case of internal data, the centroid μ is updated 1 Sum mu 2 . Repeating step 5.2 until the centroid μ 1 Sum mu 2 No change is occurring.
Step 5.4: in point mu 1 Sum mu 2 Dividing points in the coordinate system into two clusters C for centroid 1 And C 2 Wherein cluster C 1 Similarity and confidence coefficient of the inner node are compared with cluster C 2 Node similarity and confidence coefficient in the cluster C are generally higher, and cluster C is selected 1 As the similar neighbor selection range of the user i, n neighbors are selected each time, wherein the number of n is 1,3,5, 10, 20, …,140 and 150; when cluster C 1 The number of nodes in the cluster is less than n, and cluster C is selected 1 All nodes in the network are used as neighbors;
and calculating similar neighbor users of the target user Ua through the formula, and forming a similar neighbor set by collecting neighbor information.
In step 6, the following steps are included:
step 6.1: the user searches for and/or scores item Ia according to the following formula:
wherein, the liquid crystal display device comprises a liquid crystal display device,and->Historical average scores for user i and user j, respectively, n 1 Is in the neighbor of user i (S ij ×f ij ) Greater than 0.0051 neighbor number, n 2 Is in the neighbor of user i (S ij ×f ij ) Less than 0.0051 neighbor number, +.>Is the predictive scoring value, W, of user i for item alpha 1 And W is 2 Is a weight.
In step 7, the following steps are included:
step 7.1: 6.1, predicting and scoring values of similar neighbors in the similar neighbor set to the object Ia according to a formula; and recommending m items with highest scoring predictions to the user Ua.
Compared with the prior art, the invention has the advantages that: according to the invention, the similarity and the confidence coefficient between users in the recommendation system are taken as consideration factors, and k-means clustering is used for screening similar users as a selection range, so that the diversity of articles recommended by the recommendation system is improved, and the accuracy of recommendation is improved. In addition, the number of neighbors required for optimal recommendation is reduced.
Drawings
FIG. 1 is a flowchart of a recommendation system based on similarity and confidence clustering in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be further described below.
As shown in fig. 1, the present invention proposes a recommendation system based on similarity and confidence clustering, comprising the following steps:
step 1: establishing a user scoring library in a current recommendation system;
step 2: calculating the similarity of a target user Ua and other users based on the scores of the users on different articles, and generating a set Si;
step 2.1: the user scoring data is normalized with the following formula:
wherein r is i max And r i min Representing the highest score and the lowest score in the user i score record, if the highest score is equal to the lowest score, the normalized value can be assigned to be 0, eiαepsilon < -1,1 []
Step 2.2: the target user Ua searches for and/or scores item Ia;
step 2.3: for any two users i and j, the correlation calculation formula is as follows
Through the formula, integrating correlation data between the target user Ua and other users to generate a set Si;
wherein k is α Representing the degree, k, of an article i Respectively representing the degrees of users; s is S ij The influence of user i and user j is expressed and can be regarded as the similarity of user i and user j, but the similarity of the users is directional, namely the influence between two articles is different, and the user i scores the article alpha excessively, then a iα 1, otherwise 0.
Step 3: based on the relation between users in Si, taking the jointly scored articles between the target user Ua and the users as confidence coefficients to generate a set Ri;
step 3.1: for any two users i and j, the confidence coefficient calculation formula is as follows:
through the formula, integrating confidence coefficients between the target user Ua and other users to generate a set Ri;
wherein n is ij The reason for introducing this index for the number of items commonly scored by users i and j is as follows, assuming that users i and j commonly rated 100 items and users i and k commonly rated 10 items, then the similarity S between users i and j ij Will be compared with the similarity S between i and k ik And more credibility is realized.
Step 4: placing the users in a two-dimensional coordinate system based on two indexes of similarity between the users in Si and confidence coefficients between the users in Ri;
step 4.1: with similarity S between users ij Confidence coefficient n ij As an evaluation index, a two-dimensional coordinate system is established, and a user is taken as a node.
Step 5: dividing users into two types according to a k-means clustering method, and taking the type with higher similarity and confidence coefficient in the clustered two types as a similar neighbor set;
step 5.1: sample set X i ＝{(S i1 ，f i1 )，(S i2 ，f i2 )，(S i3 ，f i3 )，…，(S ij ，f ij ) Finding out the maximum similarity between user i and other usersMinimum similarity->And maximum confidence coefficient->Minimum confidence coefficient->HandleAnd->Initial centroid μ as cluster 1 Sum mu 2 ；
Step 5.2: calculating the distance between each node and centroid muThe formula is as follows：
Wherein S is μ Similarity of centroid μ, f μ Confidence coefficient for centroid μ, from node j to centroid μ 1 Sum mu 2 Distance of (2)And->The size of (2) divides the points in the coordinate system into two clusters.
Step 5.3: according to the cluster C after division 1 And C 2 In the case of internal data, the centroid μ is updated 1 Sum mu 2 . Repeating step 5.2 until the centroid μ 1 Sum mu 2 No change is occurring.
Step 5.4: in point mu 1 Sum mu 2 Dividing points in the coordinate system into two clusters C for centroid 1 And C 2 Wherein cluster C 1 Similarity and confidence coefficient of the inner node are compared with cluster C 2 Node similarity and confidence coefficient in the cluster C are generally higher, and cluster C is selected 1 As the similar neighbor selection range of the user i, n neighbors are selected each time, wherein the number of n is 1,3,5, 10, 20, …,140 and 150; when cluster C 1 The number of nodes in the cluster is less than n, and cluster C is selected 1 All nodes in the network are used as neighbors;
and calculating similar neighbor users of the target user Ua through the formula, and forming a similar neighbor set by collecting neighbor information.
Step 6: dividing neighbors in the similar neighbor set into high-reliability neighbors and low-reliability neighbors and giving different weights to the two neighbors;
step 6.1: the user searches for and/or scores item Ia according to the following formula:
wherein, the liquid crystal display device comprises a liquid crystal display device,and->Historical average scores for user i and user j, respectively, n 1 Is in the neighbor of user i (S ij ×f ij ) Greater than 0.0051 neighbor number, n 2 Is in the neighbor of user i (S ij ×f ij ) Less than 0.0051 neighbor number, +.>Is the predictive scoring value, W, of user i for item alpha 1 And W is 2 Is a weight.
Step 7: calculating the predictive score of the user on the item Ia according to the score of the similar neighbor of the user on the item Ia and the weight of the similar neighbor; and recommending the result to the user Ua according to the score.
Step 7.1: calculating a predictive score value of the similar neighbors in the similar neighbor set on the article Ia according to the formula in the step 6.1; and recommending m items with highest scoring predictions to the user Ua.
The invention will be further illustrated by the following data of specific examples:
assuming Ua is the user of the movie recommendation system, several movies in the website database were scored once, with scores distributed between 1-5.
The specific implementation steps of the algorithm are as follows:
and traversing the current user, obtaining scoring records of all movies by each user, and listing part of scoring data of the users to illustrate the implementation mode.
The similarity between users is calculated and the results are shown in the following table
The calculation of the user confidence coefficient is performed, and the table of the user confidence coefficients is as follows:
for three movies 1, movie 4, movie 6, for which user Ua did not score, a predictive score was made.
Taking a user Ua as an example, according to the relationship between the similarity and the confidence coefficient, the neighbors of the user Ua are respectively from near to far as the user U 4 User U 5 User U 2 User U 3 User U 1 . The result of k-means cluster output is user U 4 User U 5 The method comprises the steps of carrying out a first treatment on the surface of the So the neighbor selected by the user Ua is the user U 4 User U 5 。
When the neighbor selection n of the user Ua is 5 (the actual situation is generally 50-100), predicting the score of the user Ua to the movie 1 according to the formula [17], and calculating according to the formula [17], so as to obtain the possible score of the user to the movie 1 as 3.263.
Similarly, a user score for movie 4 may be calculated as 3.983; score for movie 6 as 3.623;
the prediction results are ranked, so that the order of recommending the film to the user can be obtained as follows: movie 4, movie 6, movie 1.
The target user Ua can select a proper movie according to the recommendation, and score the item after watching, so that the database can be enriched, and the personalized service of the recommendation system can be more accurate.
The foregoing is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any person skilled in the art will make any equivalent substitution or modification to the technical solution and technical content disclosed in the invention without departing from the scope of the technical solution of the invention, and the technical solution of the invention is not departing from the scope of the invention.

Claims (3)

1. A recommendation system based on similarity and confidence coefficient clustering is characterized by comprising the following steps:
step 1: establishing a user scoring library in a current recommendation system;
step 2: calculating the similarity of a target user Ua and other users based on the scores of the users on different articles, and generating a set Si;
step 3: based on the relation between users in Si, taking the jointly scored articles between the target user Ua and the users as confidence coefficients to generate a set Ri;
step 4: placing the users in a two-dimensional coordinate system based on two indexes of similarity between the users in Si and confidence coefficients between the users in Ri;
step 5: dividing users into two types according to a k-means clustering method, and taking the type with higher similarity and confidence coefficient in the clustered two types as a similar neighbor set;
step 6: dividing neighbors in the similar neighbor set into high-reliability neighbors and low-reliability neighbors and giving different weights to the two neighbors;
step 7: calculating the predictive score of the user on the item Ia according to the score of the similar neighbor of the user on the item Ia and the weight of the similar neighbor; recommending the result to the user Ua according to the score;
in step 2, the following steps are included:
step 2.1: the user scoring data is normalized with the following formula:
wherein r is i max And r i min Score record representing user iThe highest score and the lowest score in the record, if the highest score is equal to the lowest score, the normalized value can be assigned to 0, e iα ∈[-1,1]
Step 2.2: the target user Ua searches for and/or scores item Ia;
step 2.3: for any two users i and j, the correlation calculation formula is as follows
Through the formula, integrating correlation data between the target user Ua and other users to generate a set Si;
wherein k is α The degrees of the items are represented, ki represents the degrees of the users respectively; s is S ij The influence of user i and user j is expressed and can be regarded as the similarity of user i and user j, but the similarity of the users is directional, namely the influence between two articles is different, and the user i scores the article alpha excessively, then a iα 1, otherwise 0;
in step 3, the following steps are included:
step 3.1: for any two users i and j, the confidence coefficient calculation formula is as follows:
through the formula, integrating confidence coefficients between the target user Ua and other users to generate a set Ri;
wherein n is ij The reason for introducing this index for the number of items commonly scored by users i and j is as follows, assuming that users i and j commonly rated 100 items and users i and k commonly rated 10 items, then the similarity S between users i and j ij Will be compared with the similarity S between i and k ik The method has more credibility;
in step 6, the following steps are included:
step 6.1: the user searches for and/or scores item Ia according to the following formula:
wherein, the liquid crystal display device comprises a liquid crystal display device,and->Historical average scores for user i and user j, respectively, n 1 Is in the neighbor of user i (S ij ×f ij ) Greater than 0.0051 neighbor number, n 2 Is in the neighbor of user i (S ij ×f ij ) Less than 0.0051 neighbor number, +.>Is the predictive scoring value, W, of user i for item alpha 1 And W is 2 Is a weight;
in step 7, the following steps are included:
step 7.1: calculating a predictive score value of the similar neighbors in the similar neighbor set on the article Ia by a formula in step 6.1; and recommending m items with highest scoring predictions to the user Ua.
2. The recommendation system based on similarity and confidence clustering of claim 1, comprising the steps of, in step 4:
step 4.1: with similarity S between users ij Confidence coefficient n ij As an evaluation index, a two-dimensional coordinate system is established, and a user is taken as a node.
3. The recommendation system based on similarity and confidence clustering of claim 2, comprising the steps of, in step 5:
step 5.1: sample set X i ＝{S i1 ,f i1 ),(S i2 ,f i2 ),(S i3 ,f i3 ),…,(S ij ,f ij ) Finding the maximum similarity of user i with other usersDegree ofMinimum similarity->And maximum confidence coefficient->Minimum confidence coefficient->HandleAnd->Initial centroid μ as cluster 1 Sum mu 2 ；
Step 5.2: calculating the distance between each node and centroid muThe formula is as follows:
wherein S is μ Similarity of centroid μ, f μ Confidence coefficient for centroid μ, from node j to centroid μ 1 Sum mu 2 Distance of (2)Anddividing points in a coordinate system into two clusters;
step 5.3: according to the cluster C after division 1 And C 2 In the case of internal data, the centroid μ is updated 1 Sum mu 2 The method comprises the steps of carrying out a first treatment on the surface of the Repeating step 5.2 until the centroid μ 1 Sum mu 2 No change occurs;
step 5.4: in point mu 1 Sum mu 2 Dividing points in the coordinate system into two clusters C for centroid 1 And C 2 Wherein cluster C 1 Similarity and confidence coefficient of the inner node are compared with cluster C 2 Node similarity and confidence coefficient in the cluster C are generally higher, and cluster C is selected 1 As the similar neighbor selection range of the user i, n neighbors are selected each time, wherein the number of n is 1,3,5, 10, 20, …,140 and 150; when cluster C 1 The number of nodes in the cluster is less than n, and cluster C is selected 1 All nodes in the network are used as neighbors;
and calculating similar neighbor users of the target user Ua through the formula, and forming a similar neighbor set by collecting neighbor information.
CN202010006918.2A 2020-01-03 2020-01-03 Recommendation system based on similarity and confidence coefficient clustering Active CN111143699B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010006918.2A CN111143699B (en) 2020-01-03 2020-01-03 Recommendation system based on similarity and confidence coefficient clustering

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010006918.2A CN111143699B (en) 2020-01-03 2020-01-03 Recommendation system based on similarity and confidence coefficient clustering

Publications (2)

Publication Number Publication Date
CN111143699A CN111143699A (en) 2020-05-12
CN111143699B true CN111143699B (en) 2023-07-28

Family

ID=70523565

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010006918.2A Active CN111143699B (en) 2020-01-03 2020-01-03 Recommendation system based on similarity and confidence coefficient clustering

Country Status (1)

Country Link
CN (1) CN111143699B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113779421A (en) * 2021-08-13 2021-12-10 咪咕数字传媒有限公司 Association recommendation method, device, equipment and computer readable storage medium
CN116402249B (en) * 2023-03-06 2024-02-23 贝壳找房(北京)科技有限公司 Recommendation system overflow effect evaluation method, recommendation system overflow effect evaluation equipment and storage medium
CN116757794A (en) * 2023-08-17 2023-09-15 酒仙网络科技股份有限公司 Big data-based product recommendation method in wine selling applet

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6041311A (en) * 1995-06-30 2000-03-21 Microsoft Corporation Method and apparatus for item recommendation using automated collaborative filtering
CN106095974A (en) * 2016-06-20 2016-11-09 上海理工大学 Commending system score in predicting based on network structure similarity and proposed algorithm
CN108241619A (en) * 2016-12-23 2018-07-03 西北大学 A kind of recommendation method based on the more interest of user
CN109299317A (en) * 2018-11-12 2019-02-01 百度在线网络技术（北京）有限公司 Method, apparatus, storage medium and the terminal device that musical works is recommended
CN109635291A (en) * 2018-12-04 2019-04-16 重庆理工大学 A kind of recommended method of fusion score information and item contents based on coorinated training
CN109933726A (en) * 2019-03-22 2019-06-25 江西理工大学 Collaborative filtering film recommended method based on user's average weighted interest vector cluster

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6041311A (en) * 1995-06-30 2000-03-21 Microsoft Corporation Method and apparatus for item recommendation using automated collaborative filtering
CN106095974A (en) * 2016-06-20 2016-11-09 上海理工大学 Commending system score in predicting based on network structure similarity and proposed algorithm
CN108241619A (en) * 2016-12-23 2018-07-03 西北大学 A kind of recommendation method based on the more interest of user
CN109299317A (en) * 2018-11-12 2019-02-01 百度在线网络技术（北京）有限公司 Method, apparatus, storage medium and the terminal device that musical works is recommended
CN109635291A (en) * 2018-12-04 2019-04-16 重庆理工大学 A kind of recommended method of fusion score information and item contents based on coorinated training
CN109933726A (en) * 2019-03-22 2019-06-25 江西理工大学 Collaborative filtering film recommended method based on user's average weighted interest vector cluster

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
"et al. Link prediction in recommender systems with confidence measures";Zhan Su et al.;《Chaos》;20190830;第1-9页 *
"基于冷启动混合聚类的协同过滤推荐算法研究";林茂杰;《中国优秀硕士学位论文全文数据库信息科技辑》;20181215(第12期);第1-50页 *
"基于情境相似度和二次聚类的协同过滤推荐算法";蔡海尼等;《计算机科学》;20160430;第43卷(第4期);第210-213页 *
"基于深度学习的推荐算法研究";程彦龙;《中国优秀硕士学位论文全文数据库信息科技辑》;20180615(第06期);第1-50页 *
"结合用户行为和物品标签的协同过滤推荐算法";李龙生等;《计算机应用与软件》;20181231;第35卷(第6期);第248-252页 *

Also Published As

Publication number Publication date
CN111143699A (en) 2020-05-12

Similar Documents

Publication Publication Date Title
CN111143699B (en) Recommendation system based on similarity and confidence coefficient clustering
Kuzelewska Clustering algorithms in hybrid recommender system on movielens data
CN103077220B (en) A kind of personalized recommendation method based on the customer group degree of association and system
WO2010037286A1 (en) Collaborative filtering-based recommendation method and system
CN108665323B (en) Integration method for financial product recommendation system
CN105608121B (en) Personalized recommendation method and device
US10970296B2 (en) System and method for data mining and similarity estimation
CN106095974B (en) Recommendation system score prediction and recommendation algorithm based on network structure similarity
KR20150023432A (en) Method and apparatus for inferring user demographics
US11663661B2 (en) Apparatus and method for training a similarity model used to predict similarity between items
CN106127506B (en) recommendation method for solving cold start problem of commodity based on active learning
CN110334284B (en) Novel recommendation method integrating multi-factor decision making
CN106709076B (en) Social network recommendation device and method based on collaborative filtering
Gupta et al. A framework for a recommendation system based on collaborative filtering and demographics
KR101738809B1 (en) Method and System for Multi Aspect Matrix Localization by Social Matching and Communication Service in Smart Home Community based on Internet of Things
CN108717442B (en) Similar film and television recommendation method based on machine learning
WO2019184480A1 (en) Item recommendation
CN112818238A (en) Self-adaptive online recommendation method and system
CN108549729B (en) Personalized user collaborative filtering recommendation method based on coverage reduction
CN114912031A (en) Mixed recommendation method and system based on clustering and collaborative filtering
CN106055665B (en) Context-aware Web service recommendation method and system based on abnormal value elimination
Khalid et al. Scalable and practical One-Pass clustering algorithm for recommender system
CN110825971B (en) Article cold start recommendation algorithm integrating relationship mining and collaborative filtering
CN112700296B (en) Method, device, system and equipment for searching/determining business object
Wasid et al. Clustering approach for multidimensional recommender systems

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