CN114519097A - Academic paper recommendation method for heterogeneous information network enhancement - Google Patents

Academic paper recommendation method for heterogeneous information network enhancement Download PDF

Info

Publication number
CN114519097A
CN114519097A CN202210418401.3A CN202210418401A CN114519097A CN 114519097 A CN114519097 A CN 114519097A CN 202210418401 A CN202210418401 A CN 202210418401A CN 114519097 A CN114519097 A CN 114519097A
Authority
CN
China
Prior art keywords
matrix
thesis
papers
meta
features
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
CN202210418401.3A
Other languages
Chinese (zh)
Other versions
CN114519097B (en
Inventor
刘柏嵩
吴俊超
沈小烽
张雪垣
王冰源
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.)
Ningbo University
Original Assignee
Ningbo 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 Ningbo University filed Critical Ningbo University
Priority to CN202210418401.3A priority Critical patent/CN114519097B/en
Publication of CN114519097A publication Critical patent/CN114519097A/en
Application granted granted Critical
Publication of CN114519097B publication Critical patent/CN114519097B/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
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/335—Filtering based on additional data, e.g. user or group profiles
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses an academic paper recommendation method for heterogeneous information network enhancement, which comprises the following steps: step 1, constructing a heterogeneous information network, wherein the heterogeneous information network comprises 3 types of nodes including users, papers and labels, and 3 types of relations, namely, an interaction relation between users and papers, a reference relation between papers and papers, and a subordinate relation between papers and labels; step 2, learning the interactive characteristics of the user and the thesis by utilizing a matrix decomposition algorithm; step 3, inputting the interactive features into a heterogeneous graph attention network to learn high-order features of the thesis in the heterogeneous information network; step 4, utilizing the outer product to calculate and fuse the characteristics obtained by learning in the steps 2 and 3; and 5, inputting the features fused in the step 4 into a depth recommendation model prediction score. The invention solves the problem of sparse interactive data by utilizing a heterogeneous information network, and can improve the recommendation accuracy.

Description

Academic paper recommendation method for heterogeneous information network enhancement
Technical Field
The invention relates to an academic paper recommendation method, in particular to an academic paper recommendation method based on heterogeneous information network enhancement.
Background
With the explosive growth of the published amount of Academic achievements and the rapid iteration of knowledge, scientific researchers are difficult to easily find Academic papers meeting their needs, and face the problem of more and more serious Paper information overload, and Academic Paper Recommendation Systems (APRs) accurately recommend papers to researchers, and are becoming indispensable tools for researchers. Collaborative Filtering (CF) is widely used in recommendation systems, which predicts user personalized preferences by exploring historical interactions of users, however, CF cannot produce robust performance when the interaction matrix is very sparse; in recent years, many methods have been proposed to improve recommendation performance using various auxiliary information. In the paper recommendation, two kinds of auxiliary information are widely adopted: textual information and structural information.
The text information enhances the characteristics of the papers by using titles and abstracts of the papers, but the topics of the papers of the recommendation lists generated by the text information tend to be similar and lack diversity and novelty. The structural Information can be divided into a citation Network and a Heterogeneous Information Network (HIN), the citation Network reflects the relationship between one paper and another paper, but there is a cold start problem that the new paper is cited as zero; the heterogeneous information network is a method for fusing multi-source data, and the rich semantic information can effectively alleviate the problems; how to enhance the performance of paper recommendation by utilizing a heterogeneous information network has attracted the interest and attention of scientific research personnel. The basic idea of the existing thesis recommendation method based on the heterogeneous information network is to learn the characteristics of users and thesis by using a graph embedding method, calculate the scores of the users and the thesis according to the learned characteristics, and generate recommendations such as PRHNE and HGRec by ranking the scores, although the methods effectively improve the thesis recommendation performance, the following problems exist:
1) only the first-order or second-order similarity among the network nodes is considered during feature learning, and the high-order connection relation among the nodes is not excavated;
2) when the recommendation is generated, only the cosine similarity between the user and the paper features is considered, the complex interaction relationship between the user and the paper is not mined, and the enhancement effect of the heterogeneous information network on the recommendation performance cannot be fully reflected.
Therefore, there is a need to develop a new academic paper recommendation method based on heterogeneous information network enhancement to solve the existing problems.
Disclosure of Invention
The invention aims to provide an academic paper recommendation method for heterogeneous information network enhancement. The invention solves the problem of sparse interactive data by utilizing a heterogeneous information network, and can improve the recommendation accuracy.
The technical scheme of the invention is as follows: a heterogeneous information network enhanced academic paper recommendation method comprises the following steps:
step 1, constructing a heterogeneous information network, wherein the heterogeneous information network comprises 3 types of nodes including users, papers and labels, and 3 types of relations including an interaction relation between users and papers, a reference relation between papers and papers, and a subordinate relation between papers and labels;
step 2, learning the interactive characteristics of the user and the thesis by utilizing a matrix decomposition algorithm;
step 3, inputting the interactive features into a heterogeneous graph attention network to learn high-order features of the thesis in the heterogeneous information network;
step 4, utilizing the outer product to calculate and fuse the characteristics obtained by learning in the steps 2 and 3;
and 5, inputting the features fused in the step 4 into a depth recommendation model prediction score.
In the academic paper recommendation method for heterogeneous information network enhancement, the step 1 constructs a heterogeneous information network based on a citeuulike data set, the 3 types of nodes of a user, a paper and a label are respectively represented by a symbol U, P, T, and the step 1 specifically comprises the following sub-steps:
substep 1.1, converting data in the source file into a triple form: (h t r), h represents a head node id, t represents a tail node id, and r represents the relationship type between the head node h and the tail node t;
substep 1.2, establishing a relation matrix according to the triples, wherein the relation matrix comprises 3 matrixes: the method comprises the following steps of establishing a | U | × | P | interaction relation matrix between users and papers, a | P | × | P | reference relation matrix between papers and papers, and a | P | × | T | subordinate relation matrix between papers and labels, wherein the establishing process of the matrixes is as follows: firstly, initializing all 0 matrixes, then positioning the positions of matrix elements according to the triples, wherein the head nodes are matrix row numbers, the tail nodes are matrix column numbers, and the values are set to be 1;
substep 1.3, aligning the thesis id number of the relation matrix, and establishing a heterogeneous information network.
In the aforementioned academic paper recommendation method for heterogeneous information network enhancement, the step 2 specifically includes the following substeps:
substep 2.1, initializing the user interaction feature matrix at random
Figure 104061DEST_PATH_IMAGE001
And a thesis interaction feature matrix
Figure 21202DEST_PATH_IMAGE002
；
Substep 2.2 updating the matrix based on the loss function
Figure 863256DEST_PATH_IMAGE001
And
Figure 258465DEST_PATH_IMAGE002
the loss function is as follows:
Figure 72837DEST_PATH_IMAGE003
，
wherein the content of the first and second substances,
Figure 456152DEST_PATH_IMAGE004
to represent
Figure 723185DEST_PATH_IMAGE005
The user interaction feature is located at the user interface,
Figure 922085DEST_PATH_IMAGE006
to represent
Figure 653281DEST_PATH_IMAGE007
The interactive features of the paper are described,
Figure 646645DEST_PATH_IMAGE008
the dimensions of the interactive features are represented,
Figure 400974DEST_PATH_IMAGE009
representing regularization coefficients that prevent overfitting, iteratively updating the matrix
Figure 200303DEST_PATH_IMAGE001
And
Figure 723688DEST_PATH_IMAGE002
up to
Figure 153532DEST_PATH_IMAGE010
No longer decreases.
In the aforementioned academic paper recommendation method for heterogeneous information network enhancement, the step 3 specifically includes the following substeps:
substep 3.1, giving a set of paper related meta-paths
Figure 457475DEST_PATH_IMAGE011
And calculating the meta-path neighbor matrix based on the relationship matrix obtained in the substep 1.2, wherein the calculation formula is as follows:
for the meta-path PUP, it is possible to,
Figure 732598DEST_PATH_IMAGE012
，
for the meta-path PP,
Figure 376069DEST_PATH_IMAGE013
，
for the meta-path PTP the information is,
Figure 711236DEST_PATH_IMAGE014
，
wherein the content of the first and second substances,
Figure 3939DEST_PATH_IMAGE015
the transpose of the matrix is represented, the neighbor matrix obtained by the calculation needs to be converted into a binary matrix, and a threshold value is set
Figure 82753DEST_PATH_IMAGE016
When the elements in the matrix are larger than
Figure 580731DEST_PATH_IMAGE016
If so, setting the value to be 1, otherwise, setting the value to be 0, and calculating the formula as follows:
Figure 414695DEST_PATH_IMAGE017
，
wherein the content of the first and second substances,
Figure 365333DEST_PATH_IMAGE018
the values of the elements representing the ith row and the jth column of the neighbor matrix,
Figure 247838DEST_PATH_IMAGE016
to customize the threshold, by
Figure 662639DEST_PATH_IMAGE019
、
Figure 605188DEST_PATH_IMAGE020
And
Figure 43122DEST_PATH_IMAGE021
by calculation, 3 binary neighbor matrices can be obtained:
Figure 526056DEST_PATH_IMAGE022
、
Figure 998626DEST_PATH_IMAGE023
and
Figure 112075DEST_PATH_IMAGE024
the median value of the matrix is 1 to represent the neighbor relation, and the median value of the matrix is 0 without the neighbor relation;
and substep 3.2, aggregating the neighbor features based on the binary neighbor matrix, introducing node-level attention, aggregating the meaningful neighbor features to learn the target node features, wherein the calculation formula is as follows:
Figure 598158DEST_PATH_IMAGE025
，
wherein the content of the first and second substances,
Figure 822466DEST_PATH_IMAGE026
node for thesisjFor target thesis nodeiThe weight coefficient of (a) is calculated,
Figure 883963DEST_PATH_IMAGE027
in order to be a function of the power exponent,
Figure 230631DEST_PATH_IMAGE028
in order to activate the function(s),
Figure 643157DEST_PATH_IMAGE029
representing the transpose of the node attention layer query vector,
Figure 671156DEST_PATH_IMAGE030
node for thesisiThe characteristics of the interaction of (a) with (b),
Figure 915056DEST_PATH_IMAGE031
splicing operation is carried out; and aggregating the neighbor information according to the weight coefficient, wherein the calculation formula is as follows:
Figure 104729DEST_PATH_IMAGE032
，
wherein the content of the first and second substances,
Figure 4551DEST_PATH_IMAGE033
presentation paperiAccording to the coefficient
Figure 632979DEST_PATH_IMAGE026
Aggregating meta-specific paths
Figure 934647DEST_PATH_IMAGE034
Features of the neighborhood;
substep 3.3 learning different element paths through the node attention layerCharacteristic of (2)
Figure 295221DEST_PATH_IMAGE035
Introduces meta-path level attention, and merges the thesis features under different meta-pathsGFirst order features of learning articles in heterogeneous information networks
Figure 682340DEST_PATH_IMAGE036
The calculation formula is as follows:
Figure 615924DEST_PATH_IMAGE037
，
Figure 506519DEST_PATH_IMAGE038
，
wherein, the first and the second end of the pipe are connected with each other,
Figure 303574DEST_PATH_IMAGE039
respectively the weight matrix and the bias of the meta-path attention,
Figure 974727DEST_PATH_IMAGE040
as a transpose of the query vector of the meta-path attention layer,
Figure 148219DEST_PATH_IMAGE041
is as followsiThe weight coefficient corresponding to the element path,
Figure 158900DEST_PATH_IMAGE042
is the total number of meta-paths;
substep 3.4, iteratively passingLHigh-order features of layer heterogeneous graph attention network learning paper
Figure 189173DEST_PATH_IMAGE043
The calculation formula is as follows:
Figure 285305DEST_PATH_IMAGE044
，
wherein the content of the first and second substances,
Figure 996909DEST_PATH_IMAGE045
the method represents the characteristics of the thesis obtained by the attention network learning of the heterogeneous graphs of different layers.
In the aforementioned academic paper recommendation method for heterogeneous information network enhancement, step 4 is performed before step 2 and step 3 to obtain user characteristics
Figure 924414DEST_PATH_IMAGE046
Paper interaction feature
Figure 63271DEST_PATH_IMAGE047
Network node characterization of paper
Figure 646699DEST_PATH_IMAGE043
Summing the thesis interaction characteristics and the thesis network node characteristics to obtain new thesis characteristics
Figure 480705DEST_PATH_IMAGE002
=
Figure 465978DEST_PATH_IMAGE047
+
Figure 775737DEST_PATH_IMAGE043
。
In the aforementioned academic thesis recommendation method enhanced by heterogeneous information network, the step 4 is based on new thesis features
Figure 908778DEST_PATH_IMAGE002
And then the characteristics of the outer product fusion user and the thesis are utilized to obtain an interactive graph
Figure 962185DEST_PATH_IMAGE048
The calculation formula is as follows:
Figure 801965DEST_PATH_IMAGE049
，
wherein the content of the first and second substances,
Figure 282625DEST_PATH_IMAGE050
is a two-dimensional matrix, and the matrix is,
Figure 637383DEST_PATH_IMAGE051
and
Figure 494480DEST_PATH_IMAGE052
a feature vector representing a particular user and paper, the subscript of the vector being the value of where it is located.
In the aforementioned method for recommending academic papers for heterogeneous information network enhancement, the step 5 specifically includes the following substeps:
substep 5.1, constructing a neural network structure of 6 convolutional layers and 1 fully-connected layer, wherein the number of convolution kernels in each layer is 32, the size of the convolution kernels is 2 multiplied by 2, the step length is 2, the dimensionality of the fully-connected layer is 32 multiplied by 1, and then, obtaining the interactive graph in the step 4
Figure 188767DEST_PATH_IMAGE048
Inputting the prediction score of the convolutional network, and calculating according to the following formula:
Figure 902645DEST_PATH_IMAGE053
，
wherein the content of the first and second substances,
Figure 947961DEST_PATH_IMAGE054
in order to predict the score for the model,
Figure 343170DEST_PATH_IMAGE055
、
Figure 455745DEST_PATH_IMAGE056
is shown as
Figure 543787DEST_PATH_IMAGE057
The parameters of the layer convolution kernel and the bias term,
Figure 810820DEST_PATH_IMAGE058
representing convolution operations with Relu as excitationThe function of the activity is a function of the activity,
Figure 72037DEST_PATH_IMAGE059
and
Figure 475337DEST_PATH_IMAGE060
representing the weight and the bias of full connection, and flatten is matrix steering quantity operation;
substep 5.2, selecting BPR as the loss function, which optimizes the model parameters by maximizing the scoring distance of the positive and negative samples, the calculation formula is as follows:
Figure 734280DEST_PATH_IMAGE061
，
wherein the content of the first and second substances,
Figure 550926DEST_PATH_IMAGE062
a training set is represented that represents the training set,
Figure 287938DEST_PATH_IMAGE063
for the user
Figure 811323DEST_PATH_IMAGE064
Is detected in the positive sample of (a),
Figure 303484DEST_PATH_IMAGE065
for the user
Figure 279531DEST_PATH_IMAGE064
The negative sample of (a) is,
Figure 820233DEST_PATH_IMAGE028
in order to activate the function(s),
Figure 24556DEST_PATH_IMAGE066
and
Figure 359723DEST_PATH_IMAGE067
the positive and negative samples of the model prediction are scored,
Figure 88644DEST_PATH_IMAGE068
are regularization coefficients that prevent overfitting.
Compared with the prior art, the invention has the beneficial effects that: the invention provides a thought for an academic paper recommendation algorithm, which is used for solving the problems that the high-order relation between nodes and the complex interaction relation between users and papers are not mined in the conventional paper recommendation algorithm and the problem of sparse interaction data between users and papers is solved by a heterogeneous information network.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is an exemplary diagram of a heterogeneous information network;
FIG. 3 is a portion of an academic paper recommendation model diagram enhanced by heterogeneous information networks;
fig. 4 is another part of an academic paper recommendation model diagram enhanced by a heterogeneous information network.
Detailed Description
The invention is further illustrated by the following figures and examples, which are not to be construed as limiting the invention.
Example (b): a method for recommending academic papers for heterogeneous information network enhancement, the flow is shown in fig. 1.
Step 1, constructing a heterogeneous information network.
As shown in fig. 2, which is a heterogeneous information network constructed based on a citeuliuke data set, part (a) of fig. 2 represents a node type, (b) represents a heterogeneous information network, (c) represents a meta-path, and (d) represents a meta-path neighbor, and the network includes 3 types of nodes: user U, paper P and tag T, 3 relations: user paper interaction relation, reference relation among papers and paper label containing relation. The citesulide data set is a public data set suitable for the field of paper recommendation, three files, namely user.dat, positions.dat and item-tag.dat, are selected as original data, wherein the user.dat is a history clicked paper record of a user, the positions.dat is a quotation record of a paper, and the item-tag.dat is a tag record owned by the paper; the method comprises the following specific steps:
substep 1.1, converting the data in the three files into a triple form: (h t r), h represents a head node id, t represents a tail node id, and r represents the relationship type between the head node h and the tail node t;
substep 1.2, establishing a relation matrix according to the triples, wherein the relation matrix comprises 3 matrixes: the method comprises the following steps of establishing a | U | × | P | interaction relation matrix (UP) among user papers, a | P | × | P | reference relation matrix (PP) among the papers and a | P | × | T | subordinate relation matrix (PT) among paper labels, wherein the establishing process of the matrixes comprises the following steps: firstly, initializing a matrix with 0, then positioning the position of a matrix element according to the triplet, wherein the head node of the matrix element is a matrix row number, the tail node of the matrix element is a matrix column number, and the value of the matrix element is set to be 1.
Substep 1.3, aligning the thesis id number of the relation matrix, and establishing a heterogeneous information network.
And 2, learning the interactive characteristics of the user and the thesis by utilizing a matrix decomposition algorithm.
The matrix is decomposed into a classic machine learning recommendation model which takes the user paper history interaction matrix
Figure 167459DEST_PATH_IMAGE069
Decomposed into a matrix of user interaction characteristics
Figure 727753DEST_PATH_IMAGE001
And a thesis interaction feature matrix
Figure 233821DEST_PATH_IMAGE002
The method comprises the following specific steps:
substep 2.1, initializing the user interaction feature matrix at random
Figure 450038DEST_PATH_IMAGE001
And a thesis interaction feature matrix
Figure 394861DEST_PATH_IMAGE002
；
Substep 2.2 updating the matrix based on the loss function
Figure 747345DEST_PATH_IMAGE001
And
Figure 689893DEST_PATH_IMAGE002
the loss function is as follows:
Figure 190144DEST_PATH_IMAGE070
，
wherein the content of the first and second substances,
Figure 610761DEST_PATH_IMAGE004
to represent
Figure 83331DEST_PATH_IMAGE005
The user interaction feature is located at the user interface,
Figure 494983DEST_PATH_IMAGE006
to represent
Figure 420214DEST_PATH_IMAGE007
The interactive features of the paper are described,
Figure 644522DEST_PATH_IMAGE008
a dimension of the interactive feature is represented,
Figure 33915DEST_PATH_IMAGE009
representing regularization coefficients that prevent overfitting, iteratively updating the matrix
Figure 318266DEST_PATH_IMAGE001
And
Figure 730792DEST_PATH_IMAGE002
up to
Figure 555529DEST_PATH_IMAGE010
No longer decreases.
And 3, inputting the interactive features into the attention network of the heterogeneous graph to learn the high-order features of the paper in the heterogeneous information network.
The existing paper recommendation method based on the graph neural network does not consider the inherent difference of nodes and can lose important heterogeneous information. The meta path is a specific path connecting network nodes, and the meta path reflects different semantic information of the network nodes, as shown in part (c) of fig. 2, two papers can be connected through multiple meta paths, such as a paper-user-paper (PUP) and a paper-tag-paper (PTP), where the PUP indicates that the two papers are interacted by the same user, and the PTP indicates that the two papers have the same tag. The nodes of the thesis have different adjacent nodes under different meta-paths, and referring to fig. 3, on one hand, the adjacent nodes of the same meta-path may also have different importance; for example, paper-paper (PP) reflects inter-paper reference relationships, while papers may refer to papers for different aspects. On the other hand, different meta-paths may have different effects on the target node; for example, a paper-user-paper (PUP) may provide more information on learning of the characteristics of a paper than a paper-tag-paper (PTP). Considering the influence of neighbors of the paper on a target node under different meta-paths, learning the node characteristics of the paper by using a heterogeneous graph attention network, and specifically comprising the following steps:
substep 3.1, the working principle of the graph neural network is to propagate information between nodes based on neighbor matrices, giving a set of paper related element paths
Figure 737112DEST_PATH_IMAGE011
And calculating the meta-path neighbor matrix based on the relationship matrix obtained in the substep 1.2, wherein the calculation formula is as follows:
for the meta-path PUP, it is possible to,
Figure 192364DEST_PATH_IMAGE012
，
for the meta-path PP there is a path,
Figure 154503DEST_PATH_IMAGE013
，
for the meta-path PTP the information is,
Figure 720614DEST_PATH_IMAGE014
，
wherein the content of the first and second substances,
Figure 756703DEST_PATH_IMAGE015
the transposition of the expression matrix, the neighbor matrix obtained by the calculation has the problem of unbalanced data distribution, the neighbor matrix needs to be converted into a binary matrix, and a threshold value is set
Figure 943709DEST_PATH_IMAGE016
When the elements in the matrix are larger than
Figure 330828DEST_PATH_IMAGE016
If so, setting the value to be 1, otherwise, setting the value to be 0, and calculating the formula as follows:
Figure 435050DEST_PATH_IMAGE017
，
wherein the content of the first and second substances,
Figure 653542DEST_PATH_IMAGE018
the values of the elements representing the ith row and the jth column of the neighbor matrix,
Figure 450596DEST_PATH_IMAGE016
to customize the threshold, by
Figure 59432DEST_PATH_IMAGE019
、
Figure 295241DEST_PATH_IMAGE020
And
Figure 305923DEST_PATH_IMAGE021
by calculation, 3 binary neighbor matrices can be obtained:
Figure 273879DEST_PATH_IMAGE022
、
Figure 432328DEST_PATH_IMAGE023
and
Figure 878352DEST_PATH_IMAGE024
matrix ofThe median value is 1 to represent the neighbor relation, and the median value is 0 without the neighbor relation;
substep 3.2, aggregating neighbor features based on binary neighbor matrices, the traditional graph convolution network ignores different importance of the target node and its neighbor nodes, taking meta-path PP as an example, which reflects the reference relationship between two papers, referring to fig. 3, for the target node p under different meta-paths3Of two citations p2And p4The authors may refer to the two citations for different purposes; in order to distinguish the importance of different nodes to the target node, node-level attention is introduced, meaningful neighbor features are aggregated to learn the target node features, and the calculation formula is as follows:
Figure 9119DEST_PATH_IMAGE071
，
wherein the content of the first and second substances,
Figure 711758DEST_PATH_IMAGE026
node for thesisjFor target thesis nodeiThe weight coefficient of (a) is calculated,
Figure 295187DEST_PATH_IMAGE027
in order to be a function of the power exponent,
Figure 544902DEST_PATH_IMAGE028
in order to activate the function(s),
Figure 592493DEST_PATH_IMAGE029
representing the transpose of the node attention layer query vector,
Figure 902251DEST_PATH_IMAGE030
node for thesisiThe characteristics of the interaction of (a) with (b),
Figure 972976DEST_PATH_IMAGE031
splicing operation is carried out; and aggregating the neighbor information according to the weight coefficient, wherein the calculation formula is as follows:
Figure 88699DEST_PATH_IMAGE072
，
wherein the content of the first and second substances,
Figure 928479DEST_PATH_IMAGE033
presentation paperiAccording to the coefficient
Figure 409139DEST_PATH_IMAGE026
Aggregating meta-specific paths
Figure 763897DEST_PATH_IMAGE034
Features of the neighborhood;
substep 3.3, learning to obtain thesis characteristics under different element paths through the node attention layer
Figure 620995DEST_PATH_IMAGE035
Introduces meta-path level attention, and merges the thesis features under different meta-pathsGLearning first-order features of paper in heterogeneous information networks
Figure 315281DEST_PATH_IMAGE036
The calculation formula is as follows:
Figure 527694DEST_PATH_IMAGE073
，
Figure 573011DEST_PATH_IMAGE074
，
wherein the content of the first and second substances,
Figure 968220DEST_PATH_IMAGE039
respectively the weight matrix and the bias of the meta-path attention,
Figure 517013DEST_PATH_IMAGE040
as a transpose of the query vector of the meta-path attention layer,
Figure 667372DEST_PATH_IMAGE041
is as followsiStrip element pathThe corresponding weight coefficient is set to be the weight coefficient,
Figure 934405DEST_PATH_IMAGE042
is the total number of meta-paths;
substep 3.4, passing iterativelyLHigh-order features of layer heterogeneous graph attention network learning paper
Figure 867726DEST_PATH_IMAGE043
The calculation formula is as follows:
Figure 598921DEST_PATH_IMAGE075
，
wherein the content of the first and second substances,
Figure 857864DEST_PATH_IMAGE045
the method represents the characteristics of the thesis obtained by the attention network learning of the heterogeneous graphs of different layers.
And 4, calculating and fusing the characteristics obtained by learning in the steps 2 and 3 by utilizing the outer product.
User characteristics are available through steps 2 and 3
Figure 346615DEST_PATH_IMAGE046
Paper interaction feature
Figure 411523DEST_PATH_IMAGE047
Network node characterization of paper
Figure 934908DEST_PATH_IMAGE043
Summing the thesis interaction characteristics and the thesis network node characteristics to obtain new thesis characteristics
Figure 99173DEST_PATH_IMAGE002
=
Figure 904580DEST_PATH_IMAGE047
+
Figure 445283DEST_PATH_IMAGE043
And then the characteristics of the outer product fusion user and the thesis are utilized to obtain interactionDrawing (A)
Figure 823175DEST_PATH_IMAGE048
The calculation formula is as follows:
Figure 486237DEST_PATH_IMAGE076
，
wherein the content of the first and second substances,
Figure 949580DEST_PATH_IMAGE050
is a two-dimensional matrix, and the matrix is,
Figure 293973DEST_PATH_IMAGE051
and
Figure 854268DEST_PATH_IMAGE052
a feature vector representing a particular user and paper, the subscript of the vector being the value of where it is located. Compared with the calculation mode of inner product and splicing, the outer product has the following advantages: 1) the inner product calculation only obtains diagonal elements in the interactive graph, and the outer product has more modelable information and still has rich semantics even on sparse data; 2) the splicing calculation ignores the correlation among different feature dimensions, and the outer product models different dimension features, so that the features in the heterogeneous information network can be fully utilized; 3) the structure of the two-dimensional matrix is beneficial to the convolutional neural network to learn complex interactive relations, and under the same network scale, the convolutional neural network has fewer parameters than a multilayer perceptron, so that a model can be stacked in a deeper network, the enhancement effect of heterogeneous information network characteristics on recommendation is fully excavated, and stronger generalization capability is achieved.
Step 5, recommending models, namely a academic paper recommending model diagram enhanced by heterogeneous information network after splicing the models in the figures 3 and 4, wherein e in the figure 3HPoint to e in figure 4H。
In the traditional thesis recommendation method, cosine similarity prediction scoring is adopted, the scoring is sorted and a thesis list of top-k is returned; simple cosine operation cannot be used for fitting the interaction relation between the complex user and the paper, the interaction relation between the complex user and the paper can be mined by the ability of the neural network for fitting any function, a recommendation model is built by adopting a convolutional neural network, and the method comprises the following specific steps:
substep 5.1, constructing a neural network structure of 6 convolutional layers and 1 fully-connected layer, wherein the number of convolution kernels in each layer is 32, the size of the convolution kernels is 2 multiplied by 2, the step length is 2, the dimensionality of the fully-connected layer is 32 multiplied by 1, and then, obtaining the interactive graph in the step 4
Figure 360335DEST_PATH_IMAGE048
Inputting the prediction score of the convolutional network, and calculating according to the following formula:
Figure 576553DEST_PATH_IMAGE077
，
wherein the content of the first and second substances,
Figure 521375DEST_PATH_IMAGE054
in order to predict the score for the model,
Figure 873859DEST_PATH_IMAGE055
、
Figure 550828DEST_PATH_IMAGE056
is shown as
Figure 549615DEST_PATH_IMAGE057
The parameters of the layer convolution kernel and the bias term,
Figure 235811DEST_PATH_IMAGE058
representing a convolution operation, Relu being an activation function,
Figure 442801DEST_PATH_IMAGE059
and
Figure 618568DEST_PATH_IMAGE060
representing the weight and the offset of full connection, and flatten is matrix steering operation;
substep 5.2, the present invention focuses more on top-K performance, so the BPR is chosen as the loss function, which optimizes the model parameters by maximizing the scoring distance of the positive and negative samples, and the calculation formula is as follows:
Figure 543799DEST_PATH_IMAGE078
，
wherein the content of the first and second substances,
Figure 768106DEST_PATH_IMAGE062
a training set is represented that represents the training set,
Figure 157499DEST_PATH_IMAGE063
for the user
Figure 441850DEST_PATH_IMAGE064
Is detected in the positive sample of (a),
Figure 588798DEST_PATH_IMAGE065
for the user
Figure 679114DEST_PATH_IMAGE064
The negative sample of (a) is,
Figure 860696DEST_PATH_IMAGE028
in order to activate the function(s),
Figure 315948DEST_PATH_IMAGE066
and
Figure 779553DEST_PATH_IMAGE067
the positive and negative samples of the model prediction are scored,
Figure 345664DEST_PATH_IMAGE068
are regularization coefficients that prevent over-fitting.
The process of the HIN-APR algorithm of the invention is as follows:
Algorithm: HIN-APR
Input: HIN G=（V，E）; meta-path set MP; depth L;
Interaction matrix R;
Output: Predicted Function
Figure 381753DEST_PATH_IMAGE079
Initialize all parameters;
PQ=matrices-factorization（R）;
For number of training epochs do
For batch (u，v) from R do
Figure 70223DEST_PATH_IMAGE080
;
Figure 191763DEST_PATH_IMAGE081
p=P（u）;
q=Q（v）;
Figure 561564DEST_PATH_IMAGE082
;
Figure 780056DEST_PATH_IMAGE083
;
Update parameters by gradient descent;
End for
End for
Return
Figure 577111DEST_PATH_IMAGE084
;
Function
Figure 185947DEST_PATH_IMAGE085
:
Figure 93860DEST_PATH_IMAGE086
;
For l=0…L do:
For mp in MP. do:
y= node-attention(neibor(mp),
Figure 166858DEST_PATH_IMAGE087
);
Y.append(y);
End for
Figure 134814DEST_PATH_IMAGE088
= meta-path-attention (Y);
Y = [];
End for
Return
Figure 230946DEST_PATH_IMAGE089
;
in order to fully embody the advantages of the invention in academic paper recommendation, experiments are carried out on a citeuliuke academic paper recommendation data set, the data set is divided into citeuliuke-a and citeuliuke-t, the sparsity of interaction data is 0.22% and 0.07% respectively, leave-one-out is adopted in data set division, for each user in the data set, the last interaction is reserved as a test positive sample, the rest interactions are used as training positive samples, 999 papers which are not interacted before are randomly selected as test negative samples for each user, and the training negative samples are calculated according to the following formula (1: 1 are randomly chosen during the training process. Experiments are compared with 6 mainstream models at present, namely BPR-MF, Neu-MF, NGCF, HE-Rec, LGRec and CGPRec, and the Hit Rate (HR) and the normalized cumulative discount gain (NDCG) are used as evaluation indexes. The following table shows the experimental results:
Figure 503402DEST_PATH_IMAGE091
from the above table it can be seen that:
1) the performance of the HIN-APR algorithm is superior to that of other model methods on two data sets, compared with the best performance of other model methods, the HR is improved by 1.85% on average, the NDCG is improved by 3.42% on average, and the effectiveness of the HIN-APR modeling high-order connection and complex interaction on the enhancement of the recommendation performance of the thesis is verified.
2) In other model methods, the recommendation method based on the HIN is superior to the collaborative filtering method on the whole, which shows that the problem of data sparsity is effectively relieved by adding the HIN for recommendation, and the recommendation performance is improved. In the HIN-based method, indexes of all aspects of CGPRec are superior to HE-Rec and LGRec, which shows the limitation that the latter only models a low-order connection relation between nodes and the effectiveness of GCN modeling a high-order relation on recommendation; in addition, the performance of the HIN-APR is superior to that of the CGPRec because the HIN-APR introduces a node and meta-path attention layer, which can more accurately propagate neighbor information, while taking into account high-order relationships.
3) In the collaborative filtering method: compared with the recommended performances of MF-BPR, Neu-MF and NGCF, the recommended performances are improved in different degrees on two data sets, and the defects that complex interaction cannot be mined by matrix decomposition are shown; in a sparse data set citeuulike-t, promotion of Neu-MF and NGCF on MF-BPR is not as good as that in citeuulike-a, which shows that under sparse data, a collaborative filtering method faces an over-fitting problem, while HIN-APR is still greatly promoted, which shows that heterogeneous information network characteristics can be well utilized through outer product and convolution calculation, and the data sparseness problem is relieved.
The above are only preferred embodiments of the present invention, and the scope of the present invention is not limited to the above examples, and all technical solutions that fall under the spirit of the present invention belong to the scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may occur to those skilled in the art without departing from the principle of the invention, and are considered to be within the scope of the invention.

Claims (7)

1. A heterogeneous information network enhanced academic thesis recommendation method is characterized by comprising the following steps: the method comprises the following steps:
step 1, constructing a heterogeneous information network, wherein the heterogeneous information network comprises 3 types of nodes including users, papers and labels, and 3 types of relations including an interaction relation between users and papers, a reference relation between papers and papers, and a subordinate relation between papers and labels;
step 2, learning the interactive characteristics of the user and the thesis by utilizing a matrix decomposition algorithm;
step 3, inputting the interactive features into a heterogeneous graph attention network to learn high-order features of the thesis in the heterogeneous information network;
step 4, utilizing the outer product to calculate and fuse the characteristics obtained by learning in the steps 2 and 3;
and 5, inputting the features fused in the step 4 into a depth recommendation model prediction score.
2. The method of claim 1, wherein the method comprises: the step 1 is to construct a heterogeneous information network based on a citeuliuke data set, wherein 3 types of nodes, namely users, papers and labels, are respectively represented by a symbol U, P, T, and the step 1 specifically comprises the following substeps:
substep 1.1, converting data in the source file into a triple form: (h t r), h represents a head node id, t represents a tail node id, and r represents the relationship type between the head node h and the tail node t;
substep 1.2, establishing a relation matrix according to the triples, wherein the relation matrix comprises 3 matrixes: the method comprises the following steps of establishing a | U | × | P | interaction relation matrix between users and papers, a | P | × | P | reference relation matrix between papers and papers, and a | P | × | T | subordinate relation matrix between papers and labels, wherein the establishing process of the matrixes is as follows: firstly, initializing all 0 matrixes, then positioning the positions of matrix elements according to the triples, wherein the head nodes are matrix row numbers, the tail nodes are matrix column numbers, and the values are set to be 1;
substep 1.3, aligning the thesis id number of the relation matrix, and establishing a heterogeneous information network.
3. The method of claim 2, wherein the method comprises: the step 2 specifically comprises the following substeps:
substep 2.1, initializing the user interaction feature matrix at random
Figure 251293DEST_PATH_IMAGE001
And a thesis interaction feature matrix
Figure 840538DEST_PATH_IMAGE002
；
Substep 2.2 updating the matrix based on the loss function
Figure 308690DEST_PATH_IMAGE001
And
Figure 359692DEST_PATH_IMAGE002
the loss function is as follows:
Figure 315010DEST_PATH_IMAGE003
，
wherein, the first and the second end of the pipe are connected with each other,
Figure 560308DEST_PATH_IMAGE004
to represent
Figure 14292DEST_PATH_IMAGE005
The user interaction feature is located at the user interface,
Figure 885296DEST_PATH_IMAGE006
to represent
Figure 977011DEST_PATH_IMAGE007
The interactive features of the paper are described,
Figure 908058DEST_PATH_IMAGE008
the dimensions of the interactive features are represented,
Figure 849338DEST_PATH_IMAGE009
representing regularization coefficients that prevent overfitting, iteratively updating the matrix
Figure 258454DEST_PATH_IMAGE001
And
Figure 470255DEST_PATH_IMAGE002
up to
Figure 821471DEST_PATH_IMAGE010
No longer decreases.
4. The method of claim 3, wherein the method comprises the following steps: the step 3 specifically comprises the following substeps:
substep 3.1, giving a set of paper related meta-paths
Figure 779DEST_PATH_IMAGE011
And calculating the meta-path neighbor matrix based on the relationship matrix obtained in the substep 1.2, wherein the calculation formula is as follows:
for the meta-path PUP, it is possible to,
Figure 698739DEST_PATH_IMAGE012
，
for the meta-path PP,
Figure 279893DEST_PATH_IMAGE013
，
for the meta-path PTP the meta-path,
Figure 802010DEST_PATH_IMAGE014
，
wherein the content of the first and second substances,
Figure 203036DEST_PATH_IMAGE015
the transpose of the matrix is represented, the neighbor matrix obtained by the calculation needs to be converted into a binary matrix, and a threshold value is set
Figure 970266DEST_PATH_IMAGE016
When the elements in the matrix are larger than
Figure 405926DEST_PATH_IMAGE016
If so, setting the value to be 1, otherwise, setting the value to be 0, and calculating the formula as follows:
Figure 98945DEST_PATH_IMAGE017
，
wherein the content of the first and second substances,
Figure 987266DEST_PATH_IMAGE018
the values of the elements representing the ith row and the jth column of the neighbor matrix,
Figure 3591DEST_PATH_IMAGE016
to customize the threshold, by
Figure 277446DEST_PATH_IMAGE019
、
Figure 157678DEST_PATH_IMAGE020
And
Figure 18449DEST_PATH_IMAGE021
by calculation, 3 binary neighbor matrices can be obtained:
Figure 360437DEST_PATH_IMAGE022
、
Figure 505111DEST_PATH_IMAGE023
and
Figure 41396DEST_PATH_IMAGE024
the median value of the matrix is 1 to represent the neighbor relation, and the median value of the matrix is 0 without the neighbor relation;
and substep 3.2, aggregating the neighbor features based on the binary neighbor matrix, introducing node-level attention, aggregating the meaningful neighbor features to learn the target node features, wherein the calculation formula is as follows:
Figure 638731DEST_PATH_IMAGE025
，
wherein, the first and the second end of the pipe are connected with each other,
Figure 49990DEST_PATH_IMAGE026
node for thesisjFor target thesis nodeiThe weight coefficient of (a) is calculated,
Figure 49170DEST_PATH_IMAGE027
in order to be a function of the power exponent,
Figure 756357DEST_PATH_IMAGE028
in order to activate the function(s),
Figure 355834DEST_PATH_IMAGE029
representing the transpose of the node attention layer query vector,
Figure 321516DEST_PATH_IMAGE030
node for thesisiThe characteristics of the interaction of (a) with (b),
Figure 925935DEST_PATH_IMAGE031
splicing operation is carried out; and aggregating the neighbor information according to the weight coefficient, wherein the calculation formula is as follows:
Figure 53291DEST_PATH_IMAGE032
，
wherein, the first and the second end of the pipe are connected with each other,
Figure 140065DEST_PATH_IMAGE033
presentation paperiAccording to the coefficient
Figure 643858DEST_PATH_IMAGE026
Aggregating meta-specific paths
Figure 633942DEST_PATH_IMAGE034
Features of the neighborhood;
substep 3.3 learning thesis characteristics under different element paths through node attention layer
Figure 666620DEST_PATH_IMAGE035
Introduces meta-path level attention, and merges the thesis features under different meta-pathsGFirst order features of learning articles in heterogeneous information networks
Figure 975111DEST_PATH_IMAGE036
The calculation formula is as follows:
Figure 282595DEST_PATH_IMAGE037
，
Figure 596027DEST_PATH_IMAGE038
，
wherein the content of the first and second substances,
Figure 314453DEST_PATH_IMAGE039
respectively the weight matrix and the bias of the meta-path attention,
Figure 860972DEST_PATH_IMAGE040
for the transpose of the query vector of the meta-path attention layer,
Figure 457301DEST_PATH_IMAGE041
is a firstiThe weight coefficient corresponding to the element path,
Figure 405666DEST_PATH_IMAGE042
is the total number of meta-paths;
substep 3.4, iteratively passingLHigh-order features of layer heterogeneous graph attention network learning paper
Figure 560572DEST_PATH_IMAGE043
The calculation formula is as follows:
Figure 328808DEST_PATH_IMAGE044
，
wherein the content of the first and second substances,
Figure 463248DEST_PATH_IMAGE045
indicating different layersThe charting is focused on the characteristics of the thesis obtained by the network learning.
5. The method of claim 4, wherein the method comprises: before the step 4 is carried out, the user characteristics are obtained through the step 2 and the step 3
Figure 515387DEST_PATH_IMAGE046
Paper interaction feature
Figure 60769DEST_PATH_IMAGE047
Network node characterization of paper
Figure 67033DEST_PATH_IMAGE043
Summing the thesis interaction characteristics and the thesis network node characteristics to obtain new thesis characteristics
Figure 238121DEST_PATH_IMAGE002
=
Figure 895498DEST_PATH_IMAGE048
+
Figure 628093DEST_PATH_IMAGE043
。
6. The method of claim 5, wherein the method comprises: said step 4 is based on new paper features
Figure 370921DEST_PATH_IMAGE002
And then the characteristics of the outer product fusion user and the thesis are utilized to obtain an interactive graph
Figure 345699DEST_PATH_IMAGE049
The calculation formula is as follows:
Figure 608315DEST_PATH_IMAGE050
，
wherein the content of the first and second substances,
Figure 761079DEST_PATH_IMAGE051
is a two-dimensional matrix, and the matrix is,
Figure 709312DEST_PATH_IMAGE052
and
Figure 723667DEST_PATH_IMAGE053
a feature vector representing a particular user and paper, the subscript of the vector being the value of where it is located.
7. The method of claim 6, wherein the method comprises: the step 5 specifically includes the following substeps:
substep 5.1, constructing a neural network structure of 6 convolutional layers and 1 fully-connected layer, wherein the number of convolution kernels in each layer is 32, the size of the convolution kernels is 2 multiplied by 2, the step length is 2, the dimensionality of the fully-connected layer is 32 multiplied by 1, and then, obtaining the interactive graph in the step 4
Figure 90058DEST_PATH_IMAGE049
Inputting the prediction score of the convolutional network, and calculating according to the following formula:
Figure 662990DEST_PATH_IMAGE054
，
wherein the content of the first and second substances,
Figure 396722DEST_PATH_IMAGE055
in order to predict the score for the model,
Figure 932877DEST_PATH_IMAGE056
、
Figure 668621DEST_PATH_IMAGE057
is shown as
Figure 179499DEST_PATH_IMAGE058
The parameters of the layer convolution kernel and the bias term,
Figure 384215DEST_PATH_IMAGE059
representing a convolution operation, Relu being an activation function,
Figure 504487DEST_PATH_IMAGE060
and
Figure 579890DEST_PATH_IMAGE061
representing the weight and the bias of full connection, and flatten is matrix steering quantity operation;
substep 5.2, selecting BPR as the loss function, which optimizes the model parameters by maximizing the scoring distance of the positive and negative samples, the calculation formula is as follows:
Figure 261669DEST_PATH_IMAGE062
，
wherein, the first and the second end of the pipe are connected with each other,
Figure 671791DEST_PATH_IMAGE063
a training set is represented that represents the training set,
Figure 815327DEST_PATH_IMAGE064
for the user
Figure 784986DEST_PATH_IMAGE065
Is detected in the positive sample of (a),
Figure 136202DEST_PATH_IMAGE066
for the user
Figure 518773DEST_PATH_IMAGE065
The negative sample of (a) is,
Figure 747891DEST_PATH_IMAGE067
in order to activate the function(s),
Figure 312734DEST_PATH_IMAGE068
and
Figure 320004DEST_PATH_IMAGE069
the positive and negative samples of the model prediction are scored,
Figure 471762DEST_PATH_IMAGE070
are regularization coefficients that prevent over-fitting.
CN202210418401.3A 2022-04-21 2022-04-21 Academic paper recommendation method for heterogeneous information network enhancement Active CN114519097B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210418401.3A CN114519097B (en) 2022-04-21 2022-04-21 Academic paper recommendation method for heterogeneous information network enhancement

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210418401.3A CN114519097B (en) 2022-04-21 2022-04-21 Academic paper recommendation method for heterogeneous information network enhancement

Publications (2)

Publication Number Publication Date
CN114519097A true CN114519097A (en) 2022-05-20
CN114519097B CN114519097B (en) 2022-07-19

Family

ID=81600510

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210418401.3A Active CN114519097B (en) 2022-04-21 2022-04-21 Academic paper recommendation method for heterogeneous information network enhancement

Country Status (1)

Country Link
CN (1) CN114519097B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116578884A (en) * 2023-07-07 2023-08-11 北京邮电大学 Scientific research team identification method and device based on heterogeneous information network representation learning

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105843799A (en) * 2016-04-05 2016-08-10 电子科技大学 Academic paper label recommendation method based on multi-source heterogeneous information graph model
CN106815297A (en) * 2016-12-09 2017-06-09 宁波大学 A kind of academic resources recommendation service system and method
US20180240015A1 (en) * 2017-02-21 2018-08-23 Scriyb LLC Artificial cognitive declarative-based memory model to dynamically store, retrieve, and recall data derived from aggregate datasets
CN108763367A (en) * 2018-05-17 2018-11-06 南京大学 A method of scientific paper recommendation is carried out based on depth match matrix decomposition model
CN111061935A (en) * 2019-12-16 2020-04-24 北京理工大学 Science and technology writing recommendation method based on self-attention mechanism
CN111061856A (en) * 2019-06-06 2020-04-24 北京理工大学 Knowledge perception-based news recommendation method
CN111400591A (en) * 2020-03-11 2020-07-10 腾讯科技（北京）有限公司 Information recommendation method and device, electronic equipment and storage medium
CN111897974A (en) * 2020-08-12 2020-11-06 吉林大学 Heterogeneous knowledge graph learning method based on multilayer attention mechanism
CN112380434A (en) * 2020-11-16 2021-02-19 吉林大学 Interpretable recommendation system method fusing heterogeneous information network
CN113392319A (en) * 2021-05-13 2021-09-14 宁波大学 Academic paper recommendation method based on network representation and auxiliary information embedding
CN113420221A (en) * 2021-07-01 2021-09-21 宁波大学 Interpretable recommendation method integrating implicit article preference and explicit feature preference of user
CN113505294A (en) * 2021-06-15 2021-10-15 黄萌 Heterogeneous network representation recommendation algorithm fusing meta-paths

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105843799A (en) * 2016-04-05 2016-08-10 电子科技大学 Academic paper label recommendation method based on multi-source heterogeneous information graph model
CN106815297A (en) * 2016-12-09 2017-06-09 宁波大学 A kind of academic resources recommendation service system and method
US20180240015A1 (en) * 2017-02-21 2018-08-23 Scriyb LLC Artificial cognitive declarative-based memory model to dynamically store, retrieve, and recall data derived from aggregate datasets
CN108763367A (en) * 2018-05-17 2018-11-06 南京大学 A method of scientific paper recommendation is carried out based on depth match matrix decomposition model
CN111061856A (en) * 2019-06-06 2020-04-24 北京理工大学 Knowledge perception-based news recommendation method
CN111061935A (en) * 2019-12-16 2020-04-24 北京理工大学 Science and technology writing recommendation method based on self-attention mechanism
CN111400591A (en) * 2020-03-11 2020-07-10 腾讯科技（北京）有限公司 Information recommendation method and device, electronic equipment and storage medium
CN111897974A (en) * 2020-08-12 2020-11-06 吉林大学 Heterogeneous knowledge graph learning method based on multilayer attention mechanism
CN112380434A (en) * 2020-11-16 2021-02-19 吉林大学 Interpretable recommendation system method fusing heterogeneous information network
CN113392319A (en) * 2021-05-13 2021-09-14 宁波大学 Academic paper recommendation method based on network representation and auxiliary information embedding
CN113505294A (en) * 2021-06-15 2021-10-15 黄萌 Heterogeneous network representation recommendation algorithm fusing meta-paths
CN113420221A (en) * 2021-07-01 2021-09-21 宁波大学 Interpretable recommendation method integrating implicit article preference and explicit feature preference of user

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
NANA DU，等: "Recommendation of Academic Papers based on Heterogeneous Information Networks", 《2020 IEEE/ACS 17TH INTERNATIONAL CONFERENCE ON COMPUTER SYSTEMS AND APPLICATIONS (AICCSA)》 *
刘伟，等: "融合网络图模型和排序模型的论文个性化推荐算法", 《数据通信》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116578884A (en) * 2023-07-07 2023-08-11 北京邮电大学 Scientific research team identification method and device based on heterogeneous information network representation learning
CN116578884B (en) * 2023-07-07 2023-10-31 北京邮电大学 Scientific research team identification method and device based on heterogeneous information network representation learning

Also Published As

Publication number Publication date
CN114519097B (en) 2022-07-19

Similar Documents

Publication Publication Date Title
CN111523047B (en) Multi-relation collaborative filtering algorithm based on graph neural network
CN112232925A (en) Method for carrying out personalized recommendation on commodities by fusing knowledge maps
Luo et al. Personalized recommendation by matrix co-factorization with tags and time information
Xiao et al. Link prediction based on feature representation and fusion
US20230153579A1 (en) Multi-graph convolution collaborative filtering
CN115221413B (en) Sequence recommendation method and system based on interactive graph attention network
CN114519097B (en) Academic paper recommendation method for heterogeneous information network enhancement
Wang et al. CANE: community-aware network embedding via adversarial training
CN116401542A (en) Multi-intention multi-behavior decoupling recommendation method and device
CN116842260A (en) Knowledge enhancement recommendation method based on graphic neural network multi-space interaction modeling
Sridhar et al. Content-Based Movie Recommendation System Using MBO with DBN.
CN115329215A (en) Recommendation method and system based on self-adaptive dynamic knowledge graph in heterogeneous network
CN111291260A (en) Multi-information-driven approximate fusion network recommendation propagation method
CN113342994B (en) Recommendation system based on non-sampling cooperative knowledge graph network
Shu et al. Multi-task feature and structure learning for user-preference based knowledge-aware recommendation
Sangeetha et al. Predicting personalized recommendations using GNN
CN115391555A (en) User-perceived knowledge map recommendation system and method
WO2022011652A1 (en) Multi-graph convolution collaborative filtering
CN115408605A (en) Neural network recommendation method and system based on side information and attention mechanism
CN115757897A (en) Intelligent culture resource recommendation method based on knowledge graph convolution network
CN114510642A (en) Book recommendation method, system and equipment based on heterogeneous information network
Joshi et al. Interest-aware collaborative filtering recommendation model based on graph neural networks
CN114491055B (en) Recommendation method based on knowledge graph
Singh et al. Neural Network-Based Collaborative Filtering for Recommender Systems
Bougteb et al. Deep Learning-Driven Multi-Criteria Decision-Making for Effective 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