CN111079005A - Recommendation method based on article time popularity - Google Patents
Recommendation method based on article time popularity Download PDFInfo
- Publication number
- CN111079005A CN111079005A CN201911241046.1A CN201911241046A CN111079005A CN 111079005 A CN111079005 A CN 111079005A CN 201911241046 A CN201911241046 A CN 201911241046A CN 111079005 A CN111079005 A CN 111079005A
- Authority
- CN
- China
- Prior art keywords
- time
- item
- popularity
- articles
- recommendation
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 239000011159 matrix material Substances 0.000 claims abstract description 18
- 238000013468 resource allocation Methods 0.000 claims abstract description 17
- 238000012937 correction Methods 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000007423 decrease Effects 0.000 claims description 5
- 238000012986 modification Methods 0.000 claims description 2
- 230000004048 modification Effects 0.000 claims description 2
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000009792 diffusion process Methods 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
Images
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
-
- 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
- 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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a recommendation method based on article time popularity, which comprises the following steps: s1, dividing the time axis of the data set into n sections by selecting the time interval delta t, and defining the weight w of the ith time zoneiThe method comprises the steps of S2, counting distribution of articles in a time period by combining occupation ratio of the articles in the market, S3, accumulating resource allocation of each time region to obtain popularity scores of the articles α, S4, calculating a similarity matrix between the articles based on a CosRA similarity index, S5, calculating a resource allocation result according to the similarity matrix in S4, S6, adding a time factor, correcting a recommendation list of a target user i, and obtaining a final recommendation list generated by the first L articles after descending order arrangement of the recommendation list of the user i after last correction.
Description
Technical Field
The invention belongs to the technical field of article recommendation, and particularly relates to a recommendation method based on article time popularity.
Background
The rapid growth of the internet age has changed our daily lives and the way we obtain information. The information has the characteristics of high propagation speed and huge information quantity. The information coming across the sky causes a problem that the user is overloaded when acquiring the information. The problem of information overload means that people cannot quickly and accurately locate the required information from massive information. The large amount of information reduces the efficiency of obtaining information for the user, resulting in the opposite effect of providing convenience to the user. In order to solve the information overload problem, a personalized recommendation system is produced. The personalized recommendation system is used for recommending interests and love of each userAnd analyzing the similarity among the articles according to the information of good, behavior and the like, recommending the articles which are possibly interested by the user to the user, and predicting the potential behavior of the user. Meanwhile, a good personalized recommendation system is an outstanding recommendation algorithm, and many people now propose recommendation algorithms based on various technologies, such as collaborative filtering (including user-based collaborative filtering and article-based collaborative filtering)[1-2]Mass diffusion[3]And heat conduction[4]And the like.
Many excellent recommendation algorithms have been proposed based on bipartite graph calculations, which only include both user and item factors. In recent years, more and more influencing factors have been taken into account. "information aging" is an important component in the development process of information networks, but most researches focus on the structural characteristics of the networks, and neglect extremely valuable time information, which can overestimate the recommendation performance of the algorithm[5]. In the recommendation, timeliness is closely related to user satisfaction. The traditional recommendation algorithm based on the bipartite graph ignores the development rule and the market life of the articles, so that the articles which are too old are possibly recommended to the user too much, and the freshness and the chance of discovering new hobbies of the user are reduced. In addition, most of the recommendation algorithms for researching time factors mainly focus on the change of user interests, and from the perspective of articles, the development rules and market life of thought articles can have great difference according to the external environment and the characteristics of the user. The traditional recommendation algorithm has too single factor, so that the confusion of time sequence logic is easy to occur, and the experience of a user is reduced.
Reference documents:
[1]B.Sarwar,G.Karypis,J.Konstan,J.Riedl,Item-based collaborativefiltering recommendation algorithms,in:Proceedings of the 10th InternationalConference on World Wide Web,WWW’01,ACM,New York,NY,USA,2001,pp.285–295.
[2]D.Goldberg,D.Nichols,B.M.Oki,D.Terry,Using collaborative filteringto weave an information tapestry,Commun.ACM 35(12)(1992)61–70.
[3]Y.-C.Zhang,M.Medo,J.Ren,T.Zhou,T.Li,F.Yang,Recommendation modelbased on opinion diffusion,Europhys.Lett.80(6)(2007)68003.
[4]Y.-C.Zhang,M.Blattner,Y.-K.Yu,Heat conduction process on communitynetworks as a recommendation model,Phys.Rev.Lett.99(15)(2007)154301.
[5]Vidmer,A.,&Medo,Matú？.(2016).The essential role of time innetwork-based recommendation.EPL(Europhysics Letters),116(3),30007.
disclosure of Invention
The invention aims to provide a recommendation method based on the article time popularity in order to solve the problems that the conventional recommendation algorithm has single factor, is easy to generate the disorder of time sequence logic and reduces the experience of a user.
In order to achieve the purpose, the invention adopts the technical scheme that:
a recommendation method based on item temporal popularity, comprising:
s1, dividing the time axis of the data set into n sections by selecting the time interval delta t, and defining the weight w of the ith time zonei：
S2, counting the distribution of the articles in the time period according to the market proportion of the articles;
s3, accumulating the resource allocation of each time region to obtain a popularity score of the item α;
s4, calculating a similarity matrix between the articles based on the CosRA similarity index;
s5, calculating according to the similarity matrix in the S4 to obtain a resource distribution result;
and S6, adding a time factor, correcting the recommendation list of the target user i, and performing descending order arrangement on the recommendation list of the user i after the last correction to obtain the final recommendation list generated by the previous L articles.
Preferably, the specific step of step S1 includes:
the time interval delta t is selected to divide the time axis of the data set into n sections, and the time sections are marked as follows from the current time of the data set to the initial time of the data set: 0, 1, 2, 3, … …, n;
wherein, according to the exponential increasing trend, the value of the 0 zone closest to the latest time of the data set is assigned as 1, the value of the 1 zone is assigned as 1/2, the value of the 2 zone is assigned as 1/4, and so on, the value of the nth zone is assigned as 1/4
The popularity of the articles with the same degree is differentiated, and the more active the articles in the latest time zone are, the higher the score value is, the more popular the articles are in the near future; the weight of the active articles in other time zones decreases according to the distance between the time zone and the latest time, and the weight w of the ith time zoneiThe size is defined as:
preferably, the step S2, in combination with the occupancy of the item in the market, specifically includes the steps of:
after the time zone is assigned with the weight, the popularity of the items with large degree is higher than that of the items with small degree by combining the market ratio of the items, the distribution of the items in the time zone is counted, the total degree of the items α is k α, and the degree of the items appearing in the ith time zone is k αiα(ii) a The time period with higher article appearance frequency indicates that the popularity of the article in the time period is higher, and the score s of the corresponding time zone(iα)The higher:
preferably, the step S3 of accumulating the resource allocation of each time zone to obtain a popularity score for the item α comprises:
accumulating the resource allocation of each time zone to obtain a popularity score of the item α, wherein a higher score represents a higher popularity or a higher percentage of the item in the recent time zone or in the market, i.e. a higher comprehensive popularity is more popular, and conversely, a lower score represents that the item is in a non-popular stage or is out-of-date recently, and the recommendation should be suppressed, and a comprehensive popularity score s of the item ααComprises the following steps:
or
Preferably, step S4 calculates the similarity between the articles based on the CosRA similarity index and forms a similarity matrix SCosRAThe similarity between the two articles is calculated in the following way:
wherein ,
indicates the similarity of item α and item β, kiIs the degree, k, of user iαIs the degree, k, of the article αβIs the degree of the item β.
Preferably, in step S5, according to the similarity matrix in step S4, the resource allocation result is calculated as:
f′(i)＝SCosRAf(i)
wherein ,f(i)Is an n-dimensional vector for recording the initial resources, f 'for all objects of target user i'(i)The initial resource allocation result of the user after similarity calculation, i.e. recommendation list, SCosRATo represent an item-to-item similarity matrix.
Preferably, the step S6 of adding a time factor, modifying the recommendation list of the target user i, and obtaining the final recommendation list generated by the first L items after sorting the recommendation list modified by the user i in descending order includes:
adding a time factor, and correcting a recommendation list f 'of a target user i'(i)And a more reasonable recommendation list is obtained,
resource, S, representing the initial allocation of item α by target user i(α)The popularity score for item α is the final result of the revised allocation of resources to item α by user i
Recommending the user i after the last correction
And after descending order arrangement, obtaining a final recommendation list generated by the first L articles.
The recommendation method based on the article time popularity provided by the invention has the following beneficial effects:
for the traditional algorithm, the time weight can be quantified, the comprehensive popularity of the goods can be measured, the goods in the decline period can be prevented from being excessively recommended, and the goods with the periodic activity attribute can be prevented from being excessively inhibited. In addition, other limiting conditions are not needed for modifying the resource allocation of the articles in the recommendation list, and various different reference recommendation algorithms can be selected according to the actual application situation. Finally, the time factor is added to ensure the time sequence logic of the recommendation algorithm, the true performance of the algorithm can be reflected more accurately, the accuracy, diversity and novelty of the recommendation list can be effectively improved, and the problems that the conventional recommendation algorithm is too single in contained factor, the time sequence logic is easy to be confused, and the experience sense of a user is reduced are solved.
Drawings
FIG. 1 is a graph showing the actual growth rule of an item in a recommendation system based on a recommendation method for the popularity of the item in time.
Fig. 2 is a diagram of a bipartite model commonly used in a recommendation algorithm of a recommendation method based on item time popularity.
FIG. 3 is a diagram of a model for calculating the integrated popularity of an item based on a recommendation method for the temporal popularity of the item.
Fig. 4 is an overall flowchart of an item development rule-based recommendation algorithm of an item time popularity-based recommendation method.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
According to one embodiment of the application, referring to fig. 1-4, the recommendation method based on the article time popularity in the scheme comprises the following steps:
s1, dividing the time axis of the data set into n sections by selecting the time interval delta t, and defining the weight w of the ith time zonei：
S2, counting the distribution of the articles in the time period according to the market proportion of the articles;
s3, accumulating the resource allocation of each time region to obtain a popularity score of the item α;
s4, calculating a similarity matrix between the articles based on the CosRA similarity index;
s5, calculating according to the similarity matrix in the S4 to obtain a resource distribution result;
and S6, adding a time factor, correcting the recommendation list of the target user i, and performing descending order arrangement on the recommendation list of the user i after the last correction to obtain the final recommendation list generated by the previous L articles.
The invention is mainly characterized in that time factors are added into a recommendation algorithm, and data are required to be sequenced according to time sequence when data are collected and processed. The recommendation system may be represented by a bipartite graph as shown in fig. 2, where circles represent users, squares represent individual items, and the connecting lines in the graph represent the user's actions of clicking/purchasing/browsing items. The two network graph is described as G (U, O, E) where U ═ U1,u2,…,um}，O＝{o1,o2,…,on} and E＝{e1,e2,…,ezRepresents the set of user, object and user behavior, respectively.
User behavior information is represented by adjacency matrix A, whose element a is if user i has collected item α iα1 otherwise, aiα0. The recommendation algorithm finally generates a recommendation list with the length of L for the user, and uses oi LTo indicate. Time factor is given by T ═ T1,t2,…tnDenotes by t0Representing the initial time, t, of the data seteShowing the current end moment, the algorithm mainly comprises two aspects: (one) calculating a popularity score for the item. And (II) optimizing the recommendation list by combining with a benchmark recommendation algorithm.
The above steps will be described in detail below:
calculating a popularity score for an item
Step S1, selecting a time interval Δ t to divide the time axis of the data set into n segments, and sequentially marking the time segments from the current time of the data set to the initial time of the data set as: 0, 1, 2, 3, … …, n;
wherein, according to the exponential increasing trend, the value of the 0 zone closest to the latest time of the data set is assigned as 1, the value of the 1 zone is assigned as 1/2, the value of the 2 zone is assigned as 1/4, and so on, the value of the nth zone is assigned as 1/4
The popularity of the articles with the same degree is differentiated, and the more active the articles in the latest time zone are, the higher the score value is, the more popular the articles are in the near future; the weight of the active articles in other time zones decreases according to the distance between the time zone and the latest time, and the weight w of the ith time zoneiThe size is defined as:
step S2, after the time zone is assigned with weight, the popularity of the high-degree goods is higher than that of the low-degree goods by combining the market proportion of the goods, the distribution of the goods in the time zone is counted, the total number of the goods α is k α, and the number of the goods appearing in the ith time zone is kiα(ii) a The time period with higher article appearance frequency indicates that the popularity of the article in the time period is higher, and the score s of the corresponding time zone(iα)The higher:
step S3, accumulating the resource distribution of each time zone to obtain the popularity score of the item α, wherein the higher the score is, the higher the popularity of the item in the latest time zone or the higher the proportion of the item in the market is, namely, the higher the comprehensive popularity is, the more popular the item is, and conversely, the lower the score is, the more recent the item is in the non-popular stage or the item is out of date, the recommendation should be suppressed, as shown in figure 3, the weight distribution of each time zone presents the exponential growth trend of a solid line, the score of the item in each segment is accumulated, and the comprehensive popularity score S of the item α is obtainedαComprises the following steps:
or
(II) optimizing recommendation list by combining with benchmark recommendation algorithm
Step S4, the algorithm has three elements in the research process: user, item, time. When the method is combined with a benchmark recommendation algorithm based on bipartite graph calculation, the method is not influenced by other factors, and any recommendation algorithm can be selected. Calculating a similarity matrix S between the articlesCosRAWherein, the CosRA similarity index is calculated as:
wherein ,
indicates the similarity of item α and item β, kiIs the degree, k, of user iαIs the degree, k, of the article αβIs the degree of the item β.
Step S5, according to the similarity matrix in S4, the resource allocation result is calculated as follows:
f′(i)＝SCosRAf(i)
wherein ,f(i)Is an n-dimensional vector for recording the initial resources, f 'for all objects of target user i'(i)The initial resource allocation result of the user after similarity calculation, i.e. recommendation list, SCosRAIs a similarity matrix from item to item.
Step S6, adding time factor, correcting recommendation list f 'of target user i'(i)And a more reasonable recommendation list is obtained,
resource, S, representing the initial allocation of item α by target user i(α)The popularity score for item α is the final result of the revised allocation of resources to item α by user i
Recommending the user i after the last correction
And after descending order arrangement, obtaining a final recommendation list generated by the first L articles.
For the traditional algorithm, the time weight can be quantified, the comprehensive popularity of the goods can be measured, the goods in the decline period can be prevented from being excessively recommended, and the goods with the periodic activity attribute can be prevented from being excessively inhibited. In addition, other limiting conditions are not needed for modifying the resource allocation of the articles in the recommendation list, and various different reference recommendation algorithms can be selected according to the actual application situation. Finally, the time factor is added to ensure the time sequence logic of the recommendation algorithm, the true performance of the algorithm can be reflected more accurately, the accuracy, diversity and novelty of the recommendation list can be effectively improved, and the problems that the conventional recommendation algorithm is too single in contained factor, the time sequence logic is easy to be confused, and the experience sense of a user is reduced are solved.
While the embodiments of the invention have been described in detail in connection with the accompanying drawings, it is not intended to limit the scope of the invention. Various modifications and changes may be made by those skilled in the art without inventive step within the scope of the appended claims.
Claims (7)
1. A recommendation method based on item time popularity is characterized by comprising the following steps:
s1, dividing the time axis of the data set into n sections by selecting the time interval delta t, and defining the weight w of the ith time zonei：
S2, counting the distribution of the articles in the time period according to the market proportion of the articles;
s3, accumulating the resource allocation of each time region to obtain a popularity score of the item α;
s4, calculating a similarity matrix between the articles based on the CosRA similarity index;
s5, calculating according to the similarity matrix in the S4 to obtain a resource distribution result;
and S6, adding a time factor, correcting the recommendation list of the target user i, and performing descending order arrangement on the recommendation list of the user i after the last correction to obtain the final recommendation list generated by the previous L articles.
2. The item time popularity-based recommendation method according to claim 1, wherein the specific steps of the step S1 include:
the time interval delta t is selected to divide the time axis of the data set into n sections, and the time sections are marked as follows from the current time of the data set to the initial time of the data set: 0, 1, 2, 3, … …, n;
wherein, according to the exponential increasing trend, the value of the 0 zone closest to the latest time of the data set is assigned as 1, the value of the 1 zone is assigned as 1/2, the value of the 2 zone is assigned as 1/4, and so on, the value of the nth zone is assigned as 1/4
The popularity of the articles with the same degree is differentiated, and the more active the articles in the latest time zone are, the higher the score value is, the more popular the articles are in the near future; the weight of the active articles in other time zones decreases according to the distance between the time zone and the latest time, and the weight w of the ith time zoneiThe size is defined as:
3. the recommendation method based on the item time popularity according to claim 1, wherein said step S2 combines the occupancy of the item in the market, and the specific step of counting the distribution of the item over the time period comprises:
after the time zone is assigned with the weight, the popularity of the items with large degree is higher than that of the items with small degree by combining the market ratio of the items, the distribution of the items in the time zone is counted, the total degree of the items α is k α, and the degree of the items appearing in the ith time zone is k αiα(ii) a The time period with higher article appearance frequency indicates that the popularity of the article in the time period is higher, and the score s of the corresponding time zone(iα)The higher:
4. the item time popularity-based recommendation method according to claim 1, wherein said step S3 of accumulating resource allocations for each time zone, obtaining a popularity score for an item α comprises:
accumulating the resource allocation of each time zone to obtain a popularity score of the item α, wherein a higher score represents a higher popularity or a higher percentage of the item in the recent time zone or in the market, i.e. a higher comprehensive popularity is more popular, and conversely, a lower score represents that the item is in a non-popular stage or is out-of-date recently, and the recommendation should be suppressed, and a comprehensive popularity score s of the item ααComprises the following steps:
or
5. The method according to claim 1The recommendation method for the time popularity of the article is characterized in that the step S4 calculates the similarity between the article and the article based on the CosRA similarity index and forms a similarity matrix SCosRAThe similarity between the two articles is calculated in the following way:
wherein ,
indicates the similarity of item α and item β, kiIs the degree, k, of user iαIs the degree, k, of the article αβIs the degree of the item β.
6. The item time popularity-based recommendation method according to claim 1, wherein the step S5 calculates the resource allocation result according to the similarity matrix in S4 as:
f′(i)＝SCosRAf(i)
wherein ,f(i)Is an n-dimensional vector for recording the initial resources, f 'for all objects of target user i'(i)The initial resource allocation result of the user after similarity calculation, i.e. recommendation list, SCosRATo represent an item-to-item similarity matrix.
7. The item time popularity-based recommendation method according to claim 1, wherein the step S6 is implemented by adding a time factor, modifying the recommendation list of the target user i, and performing descending order arrangement on the recommendation list of the user i after the last modification to obtain the final recommendation list generated by the first L items, and includes the specific steps of:
adding a time factor, and correcting a recommendation list f 'of a target user i'(i)And a more reasonable recommendation list is obtained,
resource, S, representing the initial allocation of item α by target user i(α)The popularity score for item α is the final result of the revised allocation of resources to item α by user i
Recommending the user i after the last correction
And after descending order arrangement, obtaining a final recommendation list generated by the first L articles.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911241046.1A CN111079005B (en) | 2019-12-06 | 2019-12-06 | Recommendation method based on item time popularity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911241046.1A CN111079005B (en) | 2019-12-06 | 2019-12-06 | Recommendation method based on item time popularity |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111079005A true CN111079005A (en) | 2020-04-28 |
| CN111079005B CN111079005B (en) | 2023-05-02 |
Family
ID=70312998
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911241046.1A Active CN111079005B (en) | 2019-12-06 | 2019-12-06 | Recommendation method based on item time popularity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111079005B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112907314A (en) * | 2020-12-28 | 2021-06-04 | 桂林旅游学院 | Support Vector Machine (SVM) -based e-commerce recommendation method |
| CN113360759A (en) * | 2021-06-09 | 2021-09-07 | 南京大学 | Crowd-sourcing task recommendation method based on dual timing sequence correlation of user and project |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104866540A (en) * | 2015-05-04 | 2015-08-26 | 华中科技大学 | Personalized recommendation method based on group user behavior analysis |
| CN104899763A (en) * | 2015-05-07 | 2015-09-09 | 西安电子科技大学 | Personalized recommendation method based on bilateral diffusion of bipartite network |
| US20160048902A1 (en) * | 2005-10-11 | 2016-02-18 | Apple Inc. | Recommending Content Items |
| CN106709780A (en) * | 2016-11-14 | 2017-05-24 | 北京邮电大学 | Article recommendation method and device |
| CN106951532A (en) * | 2017-03-21 | 2017-07-14 | 深圳大学 | The evolution analysis method and device of commodity popularity |
| CN107369069A (en) * | 2017-07-07 | 2017-11-21 | 成都理工大学 | A kind of Method of Commodity Recommendation based on triangle area computation schema |
| US10157411B1 (en) * | 2014-03-13 | 2018-12-18 | Amazon Technologies, Inc. | Recommendation system that relies on RFM segmentation |
| CN109919723A (en) * | 2019-03-01 | 2019-06-21 | 西安电子科技大学 | A kind of personalized recommendation method based on user and article |
-
2019
- 2019-12-06 CN CN201911241046.1A patent/CN111079005B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160048902A1 (en) * | 2005-10-11 | 2016-02-18 | Apple Inc. | Recommending Content Items |
| US10157411B1 (en) * | 2014-03-13 | 2018-12-18 | Amazon Technologies, Inc. | Recommendation system that relies on RFM segmentation |
| CN104866540A (en) * | 2015-05-04 | 2015-08-26 | 华中科技大学 | Personalized recommendation method based on group user behavior analysis |
| CN104899763A (en) * | 2015-05-07 | 2015-09-09 | 西安电子科技大学 | Personalized recommendation method based on bilateral diffusion of bipartite network |
| CN106709780A (en) * | 2016-11-14 | 2017-05-24 | 北京邮电大学 | Article recommendation method and device |
| CN106951532A (en) * | 2017-03-21 | 2017-07-14 | 深圳大学 | The evolution analysis method and device of commodity popularity |
| CN107369069A (en) * | 2017-07-07 | 2017-11-21 | 成都理工大学 | A kind of Method of Commodity Recommendation based on triangle area computation schema |
| CN109919723A (en) * | 2019-03-01 | 2019-06-21 | 西安电子科技大学 | A kind of personalized recommendation method based on user and article |
Non-Patent Citations (8)
| Title |
|---|
| ALEXANDRE VIDMER ETC.: ""The essential role of time in network-based recommendation"", 《DOI: 10.1209/0295-5075/116/30007》 * |
| ANKUR NARANG ETC.: ""Highly scalable parallel collaborative filtering algorithm"", 《2010 INTERNATIONAL CONFERENCE ON HIGH PERFORMANCE COMPUTING》 * |
| DEMING CHU ETC.: ""Real-Time Popularity Prediction on Instagram"", 《AUSTRALASIAN DATABASE CONFERENCE-LECTURE NOTES IN COMPUTER SCIENCE》 * |
| FEI YU ETC.: "" Network-based recommendation algorithms: A review"", 《HTTPS://DOI.ORG/10.1016/J.PHYSA.2016.02.021》 * |
| JINYIN CHEN ETC.: ""Double layered recommendation algorithm based on fast density clustering: Case study on Yelp social networks dataset"", <2017 INTERNATIONAL WORKSHOP ON COMPLEX SYSTEMS AND NETWORKS> * |
| YI DING ETC.: ""Time Weight Collaborative Filtering"", 《INTERNATIONAL CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT》 * |
| 刘欣亮: ""基于用户反馈的时序二部图推荐方法"", 《河南大学学报(自然科学版)》 * |
| 贺开明: ""LBSN中基于聚类的二分图网络推荐算法"", 《科技资讯》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112907314A (en) * | 2020-12-28 | 2021-06-04 | 桂林旅游学院 | Support Vector Machine (SVM) -based e-commerce recommendation method |
| CN113360759A (en) * | 2021-06-09 | 2021-09-07 | 南京大学 | Crowd-sourcing task recommendation method based on dual timing sequence correlation of user and project |
| CN113360759B (en) * | 2021-06-09 | 2023-08-25 | 南京大学 | Crowd measurement task recommendation method based on user and project dual time sequence correlation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111079005B (en) | 2023-05-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yuan et al. | COM: a generative model for group recommendation | |
| CN103106208B (en) | A kind of streaming medium content in mobile Internet recommends method and system | |
| TWI475412B (en) | Digital content reordering method and digital content aggregator | |
| US10846613B2 (en) | System and method for measuring and predicting content dissemination in social networks | |
| CN108153791B (en) | Resource recommendation method and related device | |
| US20120185481A1 (en) | Method and Apparatus for Executing a Recommendation | |
| CN106997358A (en) | Information recommendation method and device | |
| CN104050258A (en) | Group recommendation method based on interest groups | |
| CN108460082A (en) | A kind of recommendation method and device, electronic equipment | |
| Javari et al. | Recommender systems based on collaborative filtering and resource allocation | |
| CN108876069A (en) | A kind of endowment service recommendation method | |
| CN103886001A (en) | Personalized commodity recommendation system | |
| CN109903138B (en) | Personalized commodity recommendation method | |
| CN109978657A (en) | A kind of improvement random walk chart-pattern proposed algorithm towards many intelligence platforms | |
| CN111079005A (en) | Recommendation method based on article time popularity | |
| CN106933969A (en) | Personalized recommendation system and recommendation method based on industry upstream-downstream relationship | |
| CN107122390A (en) | Recommendation system building method based on groups of users | |
| CN107679239A (en) | Recommend method in a kind of personalized community based on user behavior | |
| CN110197404A (en) | The personalized long-tail Method of Commodity Recommendation and system of popularity deviation can be reduced | |
| CN110838043A (en) | Commodity recommendation method and device | |
| CN108694234A (en) | A kind of service recommendation model based on improvement collaborative filtering | |
| CN108846042B (en) | Social network recommendation method combined with user feedback mechanism | |
| JP2014006742A (en) | Influence estimation method, device, and program | |
| CN110516709A (en) | Medium customer value method for establishing model based on hierarchical clustering | |
| Elena | The role of Instagram influencers as a source of fashion information |
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 |