CN111125537A - Session recommendation method based on graph representation - Google Patents

Session recommendation method based on graph representation Download PDF

Info

Publication number
CN111125537A
CN111125537A CN201911405693.1A CN201911405693A CN111125537A CN 111125537 A CN111125537 A CN 111125537A CN 201911405693 A CN201911405693 A CN 201911405693A CN 111125537 A CN111125537 A CN 111125537A
Authority
CN
China
Prior art keywords
user
item
graph
session
vector
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
Application number
CN201911405693.1A
Other languages
Chinese (zh)
Other versions
CN111125537B (en
Inventor
顾盼
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.)
Zhejiang Zhiduo Network Technology Co ltd
Original Assignee
China Jiliang University
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 China Jiliang University filed Critical China Jiliang University
Priority to CN201911405693.1A priority Critical patent/CN111125537B/en
Publication of CN111125537A publication Critical patent/CN111125537A/en
Application granted granted Critical
Publication of CN111125537B publication Critical patent/CN111125537B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • 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
    • G06F16/901—Indexing; Data structures therefor; Storage structures
    • G06F16/9024—Graphs; Linked lists
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0251—Targeted advertisements
    • G06Q30/0255—Targeted advertisements based on user history

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Business, Economics & Management (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Accounting & Taxation (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Finance (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a conversation recommendation method based on graph representation. The method finds the next item that the target user is most likely to interact with based on historical interaction data of the given target user. The method comprises the steps of firstly constructing a directed graph neural network of an article based on a user historical conversation sequence. The transfer relationship between the items is captured by a graph neural network. And then modeling the current conversation of the user by using a long and short memory network to extract short-term interest, and acquiring the long-term interest of the user from the latest conversation sequence of the user by using a maximum pool method. Finally, the recommendation of the item is carried out by combining the short-term interest and the long-term interest of the user. The method of the invention overcomes the defects in the prior method: (1) the complex transfer relationship of the articles in the conversation sequence cannot be captured; (2) it is not taken into account that the long-term interests of the user also change over time. Therefore, the recommendation effect implemented by the method is obviously improved compared with the prior method.

Description

Session recommendation method based on graph representation
Technical Field
The invention belongs to the technical field of internet services, and particularly relates to a conversation social contact recommendation method based on graph representation.
Background
With the development and popularity of online services, online platforms record a large amount of user behavior data. The articles which are most interesting to the user are found from the mass data and recommended to the user, so that the satisfaction degree of the user and the income of a company can be greatly improved. At this time, the recommendation system appears to be very important. The recommendation system can dig out the favorite items of the user from the mass items.
Conventional methods, such as content-based recommendation methods and collaborative filtering methods, only capture static interaction information of a user. In fact, the user's attributes and interactions are constantly updated, and this sequence data reflects the variability of the user's interests. Therefore, a recommendation system based on sequence data attracts more and more attention, and a conversation recommendation method is just a recommendation method based on sequence data. The conversation is a user interaction sequence in a period of time, the historical interaction sequence of the user is divided into a plurality of conversations, and the dynamic change of the user interest can be captured in time.
Hidasi et al applied recurrent neural network technology to conversational recommendation tasks for the first time. After that, many enhancement methods based on the recurrent neural network technology have been proposed by the scholars. However, these methods have some problems. First, it is not considered that the user's long-term interests may also change over time. Second, the transfer relationships between distant items and the spatial structure of the item network are ignored.
Disclosure of Invention
Based on the above, the invention provides a graph representation-based conversation recommendation method, which comprises the steps of giving historical interaction data of a user and an article, and digging a spatial structure relation of the article; meanwhile, the short-term interest and the long-term interest of the user are considered, and the accuracy of the recommendation method is improved.
A conversation recommendation method based on graph representation comprises the following steps:
an item graph network T is constructed based on an item set s interacted in sessions of all users:
s＝{v1，v2，…，vm}
T＝(V，E)
wherein V represents the items in the session, m represents the number of items in a certain session sequence, V represents the collection of items in the platform, and E represents the transfer relationship between the items.
Given a session s ═ v1，v2，…，vm}, any item vjIs a node of the graph T, (v)j-1，vj) For the edges of the graph network T, a consumer item v is representedj-1Thereafter the consumer item vj. And the edge value attribute of the graphIs an edge (v)j-1，vj) The number of occurrences. In order to reduce the complexity of online computation, the method adopts offline files to store the neighbor nodes of each node in the graph network T at all times.
According to the item map network T, obtaining item vector representation, and ordering:
Figure BDA0002348577130000011
where k represents the search depth in the graph network T,
Figure BDA0002348577130000012
representative node vjVector characterization at layer k, B (j) is item v at current time in item map network TjThe neighbor set of (1) is obtained by sequencing samples (sampling) from large to small according to the edge of the graph T. Function f denotes node vjIs fused to the node vjThe function f is specifically:
Figure BDA0002348577130000013
Figure BDA0002348577130000014
wherein, WkIs a graph parameter, σ is sigmoid function, AGGREGATEkThe function used is the maximum pooling (max-pooling) method:
Figure BDA0002348577130000015
wherein, max represents max operation at element level, which can effectively capture all aspects of attributes of the neighbor.
And establishing a user session representation according to the item vector representation. Target user uiA certain session of
Figure BDA0002348577130000016
Vector characterization of
Figure BDA0002348577130000017
Comprises the following steps:
Figure BDA0002348577130000018
wherein Q is a vector matrix of the item, Q (: s)i) Representing presence in session siAll of the item vectors in. The long short term memory network LSTM is a recurrent neural network method (RNN), which is a standard sequence modeling tool. The long-short term memory network LSTM can input the sessions s in sequenceiAnd outputting the current sequence vector representation.
According to user uiSession characterization to establish short-term user interest
Figure BDA0002348577130000019
Order:
Figure BDA0002348577130000021
wherein s isiFor user uiA list of the current sessions is presented,
Figure BDA0002348577130000022
for user uiCurrent session siIs used for vector characterization.
Establishing long-term user interest according to user session representation
Figure BDA0002348577130000023
Order:
Figure BDA0002348577130000024
wherein S (i) is user uiA list of recent sessions. POOLING is an element-level POOLING operation.
Merging short-term and long-term interests of users to obtain user uiUltimate interest gi：
Figure BDA0002348577130000025
Wherein the content of the first and second substances,
Figure BDA0002348577130000026
and
Figure BDA0002348577130000027
are users u respectivelyiShort-term and long-term interests of,
Figure BDA0002348577130000028
for vector splicing operations, W is a linear transformation matrix.
And recommending the item according to the final interest of the user and the item vector characterization. Article vjVector q ofjMultiplying the user interest vector by the user interest vector, and then applying a softmax function to calculate the item vjThe fraction of (c):
Figure BDA0002348577130000029
where g represents the user's interest vector,
Figure BDA00023485771300000210
representative article vjThe possibility of becoming the next interactive item. At the same time according to
Figure BDA00023485771300000211
The log-likelihood function value of (a), calculating a loss function:
Figure BDA00023485771300000212
wherein, yjRepresents vjThe one-hot code of (a) is,
Figure BDA00023485771300000213
the function is optimized using a gradient descent method.
In order to verify the technical effect of the method in sequence recommendation, the indexes Recall @20 and NDCG are observed based on the disclosed bean movie data, and the effect of the method is obviously improved compared with the latest sequence recommendation method. The invention has the following beneficial technical effects:
(1) the invention constructs a graph network about the articles by interacting the articles through all user histories, and captures the complex transfer relationships between the articles at different times.
(2) The present invention takes into account both the short-term and long-term interests of the user's dynamics, and in particular the slowly changing long-term interests of the user.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a model framework diagram of the method of the present invention.
Detailed Description
In order to more specifically describe the present invention, the following detailed description of the embodiments of the present invention is provided with reference to the accompanying drawings.
First, relevant definitions are given for the variables and formulas used.
Definition 1. U: set of users, and U ═ U1，u2，…，un}。
Definition 2. V: set of articles, and V ═ V1，v2，…，vm}。
Definition 3. T: and constructing an item graph network T based on the item sets interacted in all the user sessions.
Definition 4.B (j): item v in item graph network TjIs selected.
Definition 5.
Figure BDA00023485771300000214
User uiA session at time t, a session being a collection of items over a period of time
Figure BDA00023485771300000215
Definition 6.s (i): user uiThe set of sessions at all times is,
Figure BDA00023485771300000216
definition 7.qj: article vjIs used for vector characterization.
Definition 8.
Figure BDA00023485771300000217
User uiIs determined.
Definition 9.
Figure BDA00023485771300000218
User uiIs determined.
In conjunction with the above variable definitions, we define the final problem as: based on given user uiCurrent session
Figure BDA00023485771300000219
And a user history conversation set S (i), wherein the conversation recommendation method models the current short-term interest and the long-term interest of the user so as to recommend the user uiThe items that are most likely to be of interest in the next step are a subset of the set V.
To this end, the invention provides a novel graph representation-based session recommendation method. As shown in FIG. 2, it digs the spatial structure relationship of the items according to the historical interaction data of the user and the items. Meanwhile, the short-term interest and the long-term interest of the user are considered, and the accuracy of the recommendation method is improved.
The model is mainly composed of four modules. The first module is an item graph network (item graph), which divides the interaction history data of the user into sessions according to time intervals, constructs the item graph network according to all session lists of all users, and obtains item vector representation. The second module is session vector characterization, which uses a recurrent neural network to model the sequence of items interacting in a session. The third module is user interest modeling, capturing user short-term and long-term interests. The short-term interest is represented by a current session representation and the long-term interest is constituted by a user's most recent session sequence representation. Module one and module three are key parts of our model and contribution. Finally, in the module, the short-term and long-term interests of the user are combined to produce a recommendation.
As shown in fig. 1, one embodiment of the present invention comprises the steps of:
s100, constructing an item graph network T based on an item set S interacted in conversations of all users:
s＝{v1，v2，…，vm}
T＝(V，E)
wherein V represents the items in the session, m represents the number of items in a certain session sequence, V represents the collection of items in the platform, and E represents the transfer relationship between the items.
To capture the item and the transfer relationship between items, we use a novel approach to construct the item graph network T from all session sequences. Given a session s ═ v1，v2，…，vm}, any item vjIs a node of the graph T, (v)j-1，vj) For the edges of the graph network T, a consumer item v is representedj-1Thereafter the consumer item vj. And the edge value attribute of the graph is edge (v)j-1，vj) The number of occurrences. In order to reduce the complexity of online computation, the method adopts offline files to store the neighbor nodes of each node in the graph network T at all times.
For node vector updating, the invention firstly reduces the dimension of the node vector to d dimension, and then updates the node vector by using a graph network T, namely: s200, according to the item map network T, obtaining item vector representation, and ordering:
Figure BDA0002348577130000031
where k represents the search depth in the graph network T,
Figure BDA0002348577130000032
representative node vjVector characterization at layer k, B (j) is item v at current time in item map network TjThe neighbor set of (1) is obtained by sequencing samples (sampling) from large to small according to the edge of the graph T. Function f denotes node vjIs fused to the node vjThe function f is specifically:
Figure BDA0002348577130000033
Figure BDA0002348577130000034
wherein, WkIs a graph parameter, σ is sigmoid function, AGGREGATEkThe function used is the maximum pooling (max-pooling) method:
Figure BDA0002348577130000035
wherein, max represents max operation at element level, which can effectively capture all aspects of attributes of the neighbor.
And S300, establishing a user session representation according to the item vector representation. Target user uiA certain session of
Figure BDA0002348577130000036
Vector characterization of
Figure BDA0002348577130000037
Comprises the following steps:
Figure BDA0002348577130000038
wherein Q is a vector matrix of the item, Q (: s)i) Representing presence in session siAll of the item vectors in. The long short term memory network LSTM is a recurrent neural network method (RNN), which is a standard sequence modeling tool. The long-short term memory network LSTM can input the sessions s in sequenceiAnd outputting the current sequence vector representation.
S400, according to the user uiSession characterization to establish short-term user interestInterest of
Figure BDA0002348577130000039
Order:
Figure BDA00023485771300000310
wherein s isiFor user uiA list of the current sessions is presented,
Figure BDA00023485771300000311
for user uiCurrent session siIs used for vector characterization.
The present invention recognizes that the interests of the user are diverse and change over time. And representing the current interest of the user by adopting the session in the current time of the user, and representing the long-term interest by adopting the session in the latest time of the user. The long-term interests of the user are more stable than the short-term interests, but also vary over time.
S500, establishing long-term interest of the user according to the user session representationOrder:
Figure BDA00023485771300000313
wherein S (i) is user uiA list of recent sessions. POOLING is an element-level POOLING operation.
The POOLING operation adopted by the invention is maximum POOLING operation (max-POOLING), the POOLING operation generally comprises large POOLING operation (max-POOLING) and average POOLING operation (mean-POOLING), and experiments prove that the experimental results of the two operations have no obvious difference.
The method of representing the user's interests using the user's current session alone is very unreliable because the user occasionally clicks on an incorrect item. Therefore, the short-term interest of the user and the long-term interest of the user are considered at the same time, the recent conversation list of the user is comprehensively considered in the long-term interest, and errors caused by wrong behaviors can be corrected.Finally, the two interests are spliced, namely S600, the short-term interest and the long-term interest of the user are combined, and the user u is obtainediUltimate interest gi：
Figure BDA00023485771300000314
Wherein the content of the first and second substances,
Figure BDA00023485771300000315
and
Figure BDA00023485771300000316
are users u respectivelyiShort-term and long-term interests of,
Figure BDA00023485771300000317
for vector splicing operations, W is a linear transformation matrix.
And S700, recommending the item according to the final interest of the user and the item vector representation. Article vjVector q ofjMultiplying the user interest vector by the user interest vector, and then applying a softmax function to calculate the item vjThe fraction of (c):
Figure BDA0002348577130000041
where g represents the user's interest vector,
Figure BDA0002348577130000042
representative article vjThe possibility of becoming the next interactive item. At the same time according to
Figure BDA0002348577130000043
The log-likelihood function value of (a), calculating a loss function:
Figure BDA0002348577130000044
wherein, yjRepresents vjThe one-hot code of (a) is,
Figure BDA0002348577130000045
the function is optimized using a gradient descent method.
The foregoing description of the embodiments is provided to facilitate understanding and application of the invention by those skilled in the art. It will be readily apparent to those skilled in the art that various modifications to the above-described embodiments may be made, and the generic principles defined herein may be applied to other embodiments without the use of inventive faculty. Therefore, the present invention is not limited to the above embodiments, and those skilled in the art should make improvements and modifications to the present invention based on the disclosure of the present invention within the protection scope of the present invention.

Claims (6)

1. A conversation recommendation method based on graph representation is characterized in that:
constructing an item graph network T based on the interactive item sets s in the conversation of all users:
s＝{v1,v2,…,vm}
T＝(V,E)
wherein V represents an item in a conversation, m represents the number of items in a certain conversation sequence, V represents a set of items in a platform, and E represents a transfer relationship between the items;
according to the item map network T, obtaining item vector representation, and ordering:
Figure FDA0002348577120000011
where k represents the search depth in the graph network T,
Figure FDA0002348577120000012
representative node vjVector characterization at layer k, B (j) is item v at current time in item map network TjThe function f represents the node vjIs fused to the node vjA medium non-linear function;
establishing a user session representation according to the item vector representation; target user uiA certain session of
Figure FDA0002348577120000013
Vector characterization of
Figure FDA0002348577120000014
Comprises the following steps:
Figure FDA0002348577120000015
wherein Q is a vector matrix of the item, Q (: s)i) Representing presence in session siThe RNN is a recurrent neural network method;
according to user uiSession characterization to establish short-term user interest
Figure FDA0002348577120000016
Order:
Figure FDA0002348577120000017
wherein s isiFor user uiA list of the current sessions is presented,
Figure FDA0002348577120000018
for user uiCurrent session siThe vector characterization of (2);
establishing long-term user interest according to user session representation
Figure FDA0002348577120000019
Order:
Figure FDA00023485771200000110
wherein S (i) is user uiA recent session list; POOLING is a POOLING operation at the element level;
merging user short-term and long-term interests to obtainUser uiUltimate interest gi：
Figure FDA00023485771200000111
Wherein the content of the first and second substances,
Figure FDA00023485771200000112
and
Figure FDA00023485771200000113
are users u respectivelyi⊕ is a vector splicing operation, W is a linear transformation matrix;
recommending articles according to the final interest of the user and the article vector representation; article vjVector q ofjMultiplying the user interest vector by the user interest vector, and then applying a softmax function to calculate the item vjThe fraction of (c):
Figure FDA00023485771200000114
where g represents the user's interest vector,
Figure FDA00023485771200000115
representative article vjThe possibility of becoming the next interactive item; at the same time according to
Figure FDA00023485771200000116
The log-likelihood function value of (a), calculating a loss function:
Figure FDA00023485771200000117
wherein, yjRepresents vjThe one-hot code of (a) is,
Figure FDA00023485771200000118
the function is optimized using a gradient descent method.
2. The graph representation-based session recommendation method according to claim 1, wherein the item graph network T is constructed by:
given a session s ═ v1,v2,…,vm}, any item vjIs a node of the graph T, (v)j-1,vj) For the edges of the graph network T, a consumer item v is representedj-1Thereafter the consumer item vj(ii) a And the edge value attribute of the graph is edge (v)j-1,vj) The number of occurrences; and storing the neighbor nodes of each node in the graph network T at all times by using the offline file.
3. The graph representation-based session recommendation method according to claim 1, wherein the item v in the item graph network TjThe neighbor set b (j) of (a) is obtained by ordering samples (sampling) from large to small according to the edges in the graph T.
4. A graph-characterization-based conversational recommendation method as claimed in claim 1, wherein the Recurrent Neural Network (RNN) employs a long short term memory network (LSTM).
5. The graph representation-based session recommendation method according to claim 1, wherein the node v is represented byjIs fused to the node vjThe medium nonlinear function f is:
Figure FDA00023485771200000119
Figure FDA00023485771200000120
wherein, WkIs a graph parameter, and sigma is a sigmoid function.
6. The non-linear function f of claim 5 wherein the AGGREGATEkThe function used is the maximum pooling (max-pooling) method:
Figure FDA00023485771200000121
where max represents the max operation at the element level.
CN201911405693.1A 2019-12-31 2019-12-31 Session recommendation method based on graph representation Active CN111125537B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911405693.1A CN111125537B (en) 2019-12-31 2019-12-31 Session recommendation method based on graph representation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911405693.1A CN111125537B (en) 2019-12-31 2019-12-31 Session recommendation method based on graph representation

Publications (2)

Publication Number Publication Date
CN111125537A true CN111125537A (en) 2020-05-08
CN111125537B CN111125537B (en) 2020-12-22

Family

ID=70506298

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911405693.1A Active CN111125537B (en) 2019-12-31 2019-12-31 Session recommendation method based on graph representation

Country Status (1)

Country Link
CN (1) CN111125537B (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111581520A (en) * 2020-05-25 2020-08-25 中国人民解放军国防科技大学 Item recommendation method and system based on item importance in session
CN111581519A (en) * 2020-05-25 2020-08-25 中国人民解放军国防科技大学 Item recommendation method and system based on user intention in session
CN112256918A (en) * 2020-11-17 2021-01-22 中国计量大学 Short video click rate prediction method based on multi-mode dynamic routing
CN112256916A (en) * 2020-11-12 2021-01-22 中国计量大学 Short video click rate prediction method based on graph capsule network
CN112395504A (en) * 2020-12-01 2021-02-23 中国计量大学 Short video click rate prediction method based on sequence capsule network
CN112559904A (en) * 2020-12-16 2021-03-26 中国计量大学 Conversational social recommendation method based on door mechanism and multi-modal graph network
CN112559878A (en) * 2020-12-24 2021-03-26 山西大学 Sequence recommendation system and recommendation method based on graph neural network
CN112905886A (en) * 2021-02-22 2021-06-04 中国计量大学 Session recommendation method based on multi-interest repeated network
CN112948683A (en) * 2021-03-16 2021-06-11 山西大学 Socialized recommendation method with dynamic fusion of social information
CN113704627A (en) * 2021-09-06 2021-11-26 中国计量大学 Session recommendation method based on time interval graph
CN113704438A (en) * 2021-09-06 2021-11-26 中国计量大学 Conversation recommendation method of abnormal picture based on layered attention mechanism
CN113722599A (en) * 2021-09-06 2021-11-30 中国计量大学 Conversation recommendation method based on user long-term interest and short-term interest modeling
CN113781181A (en) * 2021-09-16 2021-12-10 中山大学 Recommendation method for assisting user interest modeling based on use of item popularity
CN115600609A (en) * 2022-10-27 2023-01-13 国电南瑞科技股份有限公司（Cn） Session recommendation method, storage medium and device based on project representation enhancement

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102402766A (en) * 2011-12-27 2012-04-04 纽海信息技术（上海）有限公司 User interest modeling method based on web page browsing
CN104246751A (en) * 2011-12-02 2014-12-24 Kddi株式会社 Recommendation device, recommendation system, recommendation method and program
CN106202205A (en) * 2016-06-28 2016-12-07 重庆大学 Migrate based on internal memory Computational frame and long and short cycle interest and the music of Fusion Model recommends method
CN108765051A (en) * 2018-04-16 2018-11-06 达而观信息科技（上海）有限公司 item recommendation method and device
CN109767301A (en) * 2019-01-14 2019-05-17 北京大学 Recommended method and system, computer installation, computer readable storage medium
US20190243749A1 (en) * 2017-11-16 2019-08-08 Servicenow, Inc. Automated diagnostic testing of databases and configurations for performance analytics visualization software

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104246751A (en) * 2011-12-02 2014-12-24 Kddi株式会社 Recommendation device, recommendation system, recommendation method and program
CN102402766A (en) * 2011-12-27 2012-04-04 纽海信息技术（上海）有限公司 User interest modeling method based on web page browsing
CN106202205A (en) * 2016-06-28 2016-12-07 重庆大学 Migrate based on internal memory Computational frame and long and short cycle interest and the music of Fusion Model recommends method
US20190243749A1 (en) * 2017-11-16 2019-08-08 Servicenow, Inc. Automated diagnostic testing of databases and configurations for performance analytics visualization software
CN108765051A (en) * 2018-04-16 2018-11-06 达而观信息科技（上海）有限公司 item recommendation method and device
CN109767301A (en) * 2019-01-14 2019-05-17 北京大学 Recommended method and system, computer installation, computer readable storage medium

Cited By (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111581519A (en) * 2020-05-25 2020-08-25 中国人民解放军国防科技大学 Item recommendation method and system based on user intention in session
CN111581520A (en) * 2020-05-25 2020-08-25 中国人民解放军国防科技大学 Item recommendation method and system based on item importance in session
CN111581519B (en) * 2020-05-25 2022-10-18 中国人民解放军国防科技大学 Item recommendation method and system based on user intention in conversation
CN111581520B (en) * 2020-05-25 2022-04-19 中国人民解放军国防科技大学 Item recommendation method and system based on item importance in session
CN112256916A (en) * 2020-11-12 2021-01-22 中国计量大学 Short video click rate prediction method based on graph capsule network
CN112256916B (en) * 2020-11-12 2021-06-18 中国计量大学 Short video click rate prediction method based on graph capsule network
CN112256918A (en) * 2020-11-17 2021-01-22 中国计量大学 Short video click rate prediction method based on multi-mode dynamic routing
CN112256918B (en) * 2020-11-17 2021-07-06 中国计量大学 Short video click rate prediction method based on multi-mode dynamic routing
CN112395504B (en) * 2020-12-01 2021-11-23 中国计量大学 Short video click rate prediction method based on sequence capsule network
CN112395504A (en) * 2020-12-01 2021-02-23 中国计量大学 Short video click rate prediction method based on sequence capsule network
CN112559904A (en) * 2020-12-16 2021-03-26 中国计量大学 Conversational social recommendation method based on door mechanism and multi-modal graph network
CN112559904B (en) * 2020-12-16 2021-10-22 中国计量大学 Conversational social recommendation method based on door mechanism and multi-modal graph network
CN112559878A (en) * 2020-12-24 2021-03-26 山西大学 Sequence recommendation system and recommendation method based on graph neural network
CN112905886B (en) * 2021-02-22 2022-02-08 中国计量大学 Session recommendation method based on multi-interest repeated network
CN112905886A (en) * 2021-02-22 2021-06-04 中国计量大学 Session recommendation method based on multi-interest repeated network
CN112948683A (en) * 2021-03-16 2021-06-11 山西大学 Socialized recommendation method with dynamic fusion of social information
CN113704627A (en) * 2021-09-06 2021-11-26 中国计量大学 Session recommendation method based on time interval graph
CN113704438A (en) * 2021-09-06 2021-11-26 中国计量大学 Conversation recommendation method of abnormal picture based on layered attention mechanism
CN113722599A (en) * 2021-09-06 2021-11-30 中国计量大学 Conversation recommendation method based on user long-term interest and short-term interest modeling
CN113704438B (en) * 2021-09-06 2022-02-22 中国计量大学 Conversation recommendation method of abnormal picture based on layered attention mechanism
CN113704627B (en) * 2021-09-06 2022-05-17 中国计量大学 Session recommendation method based on time interval graph
CN113781181A (en) * 2021-09-16 2021-12-10 中山大学 Recommendation method for assisting user interest modeling based on use of item popularity
CN113781181B (en) * 2021-09-16 2024-03-05 中山大学 Recommendation method for assisting user interest modeling based on using item popularity
CN115600609A (en) * 2022-10-27 2023-01-13 国电南瑞科技股份有限公司（Cn） Session recommendation method, storage medium and device based on project representation enhancement
CN115600609B (en) * 2022-10-27 2024-05-14 国电南瑞科技股份有限公司 Session recommendation method, storage medium and device based on project representation enhancement

Also Published As

Publication number Publication date
CN111125537B (en) 2020-12-22

Similar Documents

Publication Publication Date Title
CN111125537B (en) Session recommendation method based on graph representation
CN111222054B (en) Session social contact recommendation method based on context neighbor relation modeling
CN109767301B (en) Recommendation method and system, computer device and computer readable storage medium
CN111460331B (en) Conversation recommendation method based on space-time diagram
CN110910218A (en) Multi-behavior migration recommendation method based on deep learning
CN112256980A (en) Dynamic graph attention network-based multi-relation collaborative filtering recommendation
CN114036406A (en) Recommendation method and system based on graph contrast learning and social network enhancement
CN112256916B (en) Short video click rate prediction method based on graph capsule network
CN112559878B (en) Sequence recommendation system and recommendation method based on graph neural network
CN112819575B (en) Session recommendation method considering repeated purchasing behavior
CN114265986A (en) Information pushing method and system fusing knowledge graph structure and path semantics
CN112528165A (en) Session social recommendation method based on dynamic routing graph network
CN113641811B (en) Session recommendation method, system, equipment and storage medium for promoting purchasing behavior
CN113704438B (en) Conversation recommendation method of abnormal picture based on layered attention mechanism
CN112559904B (en) Conversational social recommendation method based on door mechanism and multi-modal graph network
CN114282077A (en) Session recommendation method and system based on session data
CN112256918B (en) Short video click rate prediction method based on multi-mode dynamic routing
US20240037133A1 (en) Method and apparatus for recommending cold start object, computer device, and storage medium
CN116452293A (en) Deep learning recommendation method and system integrating audience characteristics of articles
CN113704439B (en) Conversation recommendation method based on multi-source information heteromorphic graph
CN114428912A (en) Session recommendation method based on capturing long-term and short-term interest heterogeneous hypergraph of user
CN112307257A (en) Short video click rate prediction method based on multi-information node graph network
CN114528490A (en) Self-supervision sequence recommendation method based on long-term and short-term interests of user
CN113744023A (en) Dual-channel collaborative filtering recommendation method based on graph convolution network
CN114036400B (en) Hypergraph-based collaborative session recommendation method

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
TR01 Transfer of patent right

Effective date of registration: 20230921

Address after: Room 407-10, floor 4, building 2, Haichuang science and technology center, Cangqian street, Yuhang District, Hangzhou City, Zhejiang Province, 311100

Patentee after: Zhejiang Zhiduo Network Technology Co.,Ltd.

Address before: 310018, No. 258, source street, Xiasha Higher Education Park, Hangzhou, Zhejiang

Patentee before: China Jiliang University

TR01 Transfer of patent right