CN109447261A - A method of the network representation study based on multistage neighbouring similarity - Google Patents
A method of the network representation study based on multistage neighbouring similarity Download PDFInfo
- Publication number
- CN109447261A CN109447261A CN201811175451.3A CN201811175451A CN109447261A CN 109447261 A CN109447261 A CN 109447261A CN 201811175451 A CN201811175451 A CN 201811175451A CN 109447261 A CN109447261 A CN 109447261A
- Authority
- CN
- China
- Prior art keywords
- node
- similarity
- neighbouring
- bac
- order
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 18
- 238000004891 communication Methods 0.000 claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims abstract description 9
- 239000013604 expression vector Substances 0.000 claims abstract description 6
- 239000013598 vector Substances 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 2
- 230000009191 jumping Effects 0.000 claims description 2
- 238000012916 structural analysis Methods 0.000 abstract description 2
- 238000011160 research Methods 0.000 description 7
- 230000001568 sexual effect Effects 0.000 description 2
- 241000282326 Felis catus Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A kind of method that the present invention proposes network representation study based on multistage neighbouring similarity.It only considered relationship of the single order adjacent to similarity and second order adjacent to similarity compared to traditional network representation based on structural analysis, high-order of the emphasis of the present invention between node is adjacent to similarity modeling, the calculation method of different classes of indirect neighbor similarity is separately designed, particularly in view of information can decay during Internet communication with the increase of distance, therefore, the present invention can predict the different neighbor nodes of present node, it more accurately finds and the maximum adjacent node of the degree of association of destination node, it is semantic more abundant so as to obtain, expression vector with higher reliability and authenticity.
Description
(1) technical field
The present invention relates to Complex Networks Analysis fields, and in particular to a kind of net list dendrography based on multistage neighbouring similarity
The method of habit.
(2) background technique
In daily life, network data is ubiquitous, for example, thousands of Website page is constituted on internet
The network of web page interlinkage, microblogging and Twitter etc. constitute the interpersonal network in people's social activity, Jingdone district and day cat
Etc. constitute user shopping network.Therefore, the presence of information network has become the most common carrier and form in our lives,
There is important learning value and application value to the research of information network.
Network representation study is also known as internet startup disk or figure insertion, is substantially with a low-dimensional, dense vector removes table
Show the node in network, the structure of the vector energy reaction network can be used for the cluster of network node, and classification task can also answer
In the tasks such as prediction and reconstruct with relationship in a network.With the development of machine learning and nerual network technique, network representation
The research of study is also more and more important, and find suitable network representation data as input is for the parameter learning of neural network
It is essential.
Currently, the research of network representation study can be divided into two kinds, one is the researchs based on network structure, such as
The methods of DeepWalk, Line, node2vec, one is the researchs based on complex network attribute, in conjunction with social networks without mark
Degree row, community cultule etc..In addition, there are also the network representation study based on content, the network representation study based on temporal dynamic property, with
And the network representation study in heterogeneous network also obtains higher and higher research temperature in recent years.But these above-mentioned methods are all
Based on the single order similitude or second order similitude between node, few the case where being discussed again to higher order similitude.Although node
Between indirect neighbor relationship can decay with the increase of node path length, but the indirect similarity between node to research node
Between convergent sexual intercourse be also highly important.
(3) summary of the invention
For disadvantages mentioned above and deficiency, the purpose of the present invention is to provide a kind of net lists based on multistage neighbouring similarity
The method that dendrography is practised.
To achieve the above objectives, technical solution of the present invention includes are as follows:
A kind of network representation learning method based on multistage neighbouring similarity, comprising:
1) network topology is abstracted and turns to the network structure comprising multiple nodes and side.
2) according to the proximity relations between nodes, the modeling of similarity is carried out to each pair of node.
3) all proximity relations of comprehensive each node pair calculate comprehensive neighbouring similarity.
4) each node combines the similarity relationship of all context nodes to be obtained according to the algorithm of network structure model
The high-order of final knot vector indicates.
The step 1) includes: to set non-directed graph G=(V, E) to indicate network topology structure, and wherein V indicates user node, E table
Show connection relationship.
The step 2) includes: that proximity relations between node pair includes two kinds: being directly adjacent to and indirect neighbor.It specifically includes
Following steps:
2.1. it is directly adjacent to based on single order similitude
If (B, C) is destination node pair, if B and C, which have, directly connects side, the path length between B, C is 1, path length
Similarity between 1 node corresponds to single order similarity.
The neighbouring calculation method of single order:
PBC=UB·UC
2.2. the indirect neighbor based on second order similitude
What network single order similitude was portrayed is to have the local feature for directly connecting side in network, and the information of description has unilateral
Property.And the company side in network topology has sparsity feature, only is difficult to meet true node with the relationship for directly connecting side convergent
Sexual intercourse.Then, there is the proposition of second order similarity.
If (B, C) is destination node pair, if not connecting side between B and C directly, but there are public single order neighbours between B and C
The shortest path length of node, i.e. B and C is 2, then is connected at B with C for second order, and the similarity between B and C is second order similarity.
The neighbouring calculation method of second order:
PBAC=WAB(UA·UB)+WAC(UA·UC)+αLBAC+βMA
2.3. the indirect neighbor based on high-order similitude
In large scale network, the proximity relations between node only study to second order be it is far from being enough, be on the one hand still
There is the problem of sparsity, on the other hand, the syntople that path length is 2 is difficult to portray global characteristics, and then, this patent mentions
The depicting method of high-order proximity relations is gone out.
If (B, C) is destination node pair, if shortest path length between B and C is k (k >=3), B and C are that k rank is adjacent
Closely.
The neighbouring calculation method of three ranks:
PBAC=WAB*(UB|UA)+WAC*(UA·UC)+αLBAC+βMA
The neighbouring calculation method of K rank: (k >=4)
PBAC=WAB(UB|UA)+WAC(UC|UA)+αLBAC+βMA
The step 3) includes: to set (B, C) as destination node pair, and there are many proximity relations, NN between B and Ck(B, C) generation
The k rank proximity of table B and C, the then overall proximity between B and C
The step 4) includes: to set each node A to have a context node set SA, SAIn element with key-value pair
Form exist, the key of element is the context adjacent node of node A, and the value of element is the degree of association of A and the context node.
Parameter information of the invention has:
Compared with the prior art, the invention has the benefit that
The method for the network representation study based on multistage neighbouring similarity that the present invention provides a kind of, compared to traditional base
In the network representation of structural analysis only consider single order adjacent to similarity and second order adjacent to the relationship of similarity, the present invention is between node
High-order adjacent to similarity modeling, particularly in view of information can decline during Internet communication with the increase of distance
Subtract, therefore, can more accurately find with the maximum adjacent node of the degree of association of destination node, so as to obtain it is semantic more
It is abundant, the expression vector with higher reliability and authenticity.
(4) Detailed description of the invention
Fig. 1 is that the present invention is based on the control flow block diagrams of the network representation learning method of multistage neighbouring similarity.
(5) specific embodiment
As shown in Figure 1, the invention discloses a kind of network representation learning methods based on multistage neighbouring similarity, comprising:
(1) true social network structure is abstracted as non-directed graph G (V, E), wherein V indicates that user node, E indicate user
Between pay close attention to and be concerned relationship.
(2) a node A in network is taken out, finds out the adjacent node for being no more than k with its step-length, and these nodes are put
Enter the context node collection S of AAIn.Each node has a corresponding context node set, and form is as follows: SA=[B:
NNAB],[C:NNAC],....,[Q:NNAQ],
Wherein [] indicates that context node element, the node in element exist in the form of key-value pair, and the key of element indicates
The title of context node, the value of element are the degree of association of context node and origin node.
(3) the context node collection S of node A is initializedAIn each node and A the degree of association be 1.
(4) S is taken outAIn any one node elements B, if path length between node B and node A is 1, i.e., directly
It is adjacent, it is calculate by the following formula the degree of association.
PAB=UA·UB
Update SAIn corresponding context node value.(note: update is not assignment, is weighting)
(5) step (4) are repeated, finishes S until updatingAIn all nodes adjacent with node A single order value.
(6) from set SAIn arbitrarily take out two nodes B and C (second traverse), judge respectively node A and node B and
Adjacency between node C, if B and C be it is neighbouring by the second order of A, then follow the steps (7)；If A and B (C) are single order neighbours
Closely, A and C (B) is that k rank is neighbouring, (k >=2), then follow the steps (8)；If A and B is that k1 rank is neighbouring, A and C is that k2 rank is neighbouring,
(k1, k2 >=2), then follow the steps (9).
(7) communication node A has been directly connected to target node b and C, i.e. node B and node C are that second order is neighbouring, communication node A
So that the probability NN communicated between B and C2It is shown below:
PBAC=WAB*(UA·UB)+WAC*(UA·UC)+αLBAC+βMA
Wherein UjIt is indicated for the one-hot vector of node j, WABFor the weight between node A and B, LikjIt indicates through communication section
Path distance between point k connected node i and j, α are propagation attenuation coefficient, MAFor the influence power of node A, node A can be used
Degree indicate that the node influence power that β is communication node A is to the biasing coefficient of the BC degree of association.
(8) communication node A is 1 and k at a distance from target node b and C, (k >=2), communication node A makes between B and C
The probability NN of communicationk+1It is shown below
PBAC=WAB*(UB·UA)+WAC*(UA|UC)+αLBAC+βMA
(9) communication node A is k at a distance from target node b1, it is k at a distance from destination node C2, communication node A makes B
The probability NN communicated between Ck1+k2It is shown below
PBAC=WAB*(UB|UA)+WAC*(UC|UA)+αLBAC+βMA
(10) S is updated by the indirect neighbor value that (7) or (8) or (9) are calculatedBAnd SCIn key be C and member that key is B
Element value.(note: update is not assignment, is weighting)
(11)SAIn node whether by all secondary traversals, if so, (12) are thened follow the steps, if it is not, then jumping
It returns step (6).
(12) it selects and unduplicated destination node in (2), execution step (2), until all nodes in network topology
All it has been expressed study.
(13) representation method for passing through skip-gram term vector, by origin node A and set of context SARespectively as nerve
Network is output and input, and the parameter learnt is the expression vector of node A.
(14) the expression vector of all nodes in network is obtained by step (13) traversal.
Claims (1)
1. the method for the network representation study based on multistage neighbouring similarity that there is provided herein a kind of, it is characterised in that:
1) actual social networks topological abstract is turned into non-directed graph, the point in non-directed graph indicates user's section in social networks
Point, the side in non-directed graph indicate relationship between the user in social networks.
2) according to the syntople between nodes, similarity modeling is carried out to each pair of node, wherein similarity includes node
Between direct similarity and indirect similarity.
3) all of its neighbor relationship of comprehensive each node pair calculates comprehensive neighbouring similarity, is stored in the context section of each node
In point set.
4) all direct or indirect context nodes of conformity goal node export target section by the method for skip-gram
The vector of point indicates.
The step 1) includes: to set non-directed graph G=(V, E) to indicate network topology structure, and wherein V indicates that user node, E indicate to connect
Connect relationship.
The step 2) includes: that proximity relations between node pair includes two kinds: being directly adjacent to and indirect neighbor.It specifically includes following
Step:
2.1. it is directly adjacent to based on single order similitude
If (B, C) is destination node pair, if B and C, which have, directly connects side, the path length between B, C is 1, and path length is 1
Similarity between node corresponds to single order similarity.
The neighbouring calculation method of single order:
PBC=UB·UC
2.2. the indirect neighbor based on second order similitude
What network single order similarity was portrayed is the local feature having between the node for directly connecting side in network, and the information of description has piece
Face property.And the company side in network topology has sparsity feature, only is difficult to meet true node with the relationship for directly connecting side and
Homosexuality.Then, there is the proposition of second order similarity.
If (B, C) is destination node pair, if not connecting side between B and C directly, but there is public single order neighbours section between B and C
The shortest path length of point, i.e. B and C is 2, then B is referred to as that second order is connected with C, and the similarity between B and C is second order similarity.
The neighbouring calculation method of second order:
PBAC=WAB(UA·UB)+WAC(UA·UC)+αLBAC+βMA
2.3. the indirect neighbor based on high-order similitude
In large scale network, on the one hand it is still to have that it is far from being enough that the proximity relations between node, which is only studied to second order,
The problem of sparsity, on the other hand, the syntople that path length is 2, are difficult to portray global characteristics, and then, this patent proposes
The depicting method of high-order proximity relations.
If (B, C) is destination node pair, if shortest path length between B and C is k (k >=3), B and C is neighbouring for k rank.
The neighbouring calculation method of three ranks:
PBAC=WAB*(UB|UA)+WAC*(UA·UC)+αLBAC+βMA
The neighbouring calculation method of K rank: (k >=4)
PBAC=WAB(UB|UA)+WAC(UC|UA)+αLBAC+βMA
The step 3) includes: to set (B, C) as destination node pair, and there are many proximity relations, NN between B and Ck(B, C) represents B and C
K rank proximity, then the overall proximity between B and C
The step 4) includes: to set each node A to have a context node set SA, SAIn element with the shape of key-value pair
Formula exists, and the key of element is the context adjacent node of node A, the degree of association of the value of element between A and the context node.This
It is related to following parameter information in invention:
Specific implementation steps are as follows:
(1) true social network structure is abstracted as non-directed graph G (V, E), wherein V indicates that user node, E indicate to close between user
Infuse and be concerned relationship.
(2) a node A in network is taken out, finds out the adjacent node for being no more than k with its step-length, and these nodes are put into A
Context node collection SAIn.Each node has a corresponding context node set, and form is as follows:
SA={ [B:NNAB],[C:NNAC],....,[Q:NNAQ],
Wherein, [] indicates that context node element, the node in element exist in the form of key-value pair, and the key of element indicates up and down
The title of literary node, the value of element are the degree of association of context node and origin node.
(3) the context node collection S of node A is initializedAIn each node and A the degree of association be 1.
(4) S is taken outAIn any one node elements B, if path length between node B and node A is 1, i.e. direct neighbor,
It is calculate by the following formula the degree of association.
PAB=UA·UB
Update SAIn corresponding context node value.(note: update is not assignment, is weighting)
(5) step (4) are repeated, finishes S until updatingAIn all nodes adjacent with node A single order value.
(6) from set SAIn arbitrarily take out two nodes B and C (second traverse), judge node A and node B and node C respectively
Between adjacency, if B and C be it is neighbouring by the second order of A, then follow the steps (7)；If A and B (C) is that single order is neighbouring, A and C
It (B) is that k rank is neighbouring, (k >=2), then follow the steps (8)；If A and B is that k1 rank is neighbouring, A and C is that k2 rank is neighbouring, (k1, k2 >=
2) (9), are thened follow the steps.
(7) communication node A has been directly connected to target node b and C, i.e. node B and node C are that second order is neighbouring, and communication node A makes
The probability NN communicated between B and C2It is shown below:
PBAC=WAB*(UA·UB)+WAC*(UA·UC)+αLBAC+βMA
Wherein UjIt is indicated for the one-hot vector of node j, WABFor the weight between node A and B, LikjIt indicates through communication node k
Path distance between connected node i and j, α are propagation attenuation coefficient, MAIt, can be with node A's for the influence power of node A
Degree expression, biasing coefficient of the node influence power to the BC degree of association that β is communication node A.
(8) communication node A is 1 and k at a distance from target node b and C, (k >=2), communication node A to communicate between B and C
Probability NNk+1It is shown below
PBAC=WAB*(UB·UA)+WAC*(UA|UC)+αLBAC+βMA
(9) communication node A is k at a distance from target node b1, it is k at a distance from destination node C2, communication node A makes B and C
Between the probability that communicatesIt is shown below
PBAC=WAB*(UB|UA)+WAC*(UC|UA)+αLBAC+βMA
(10) S is updated by the indirect neighbor value that (7) or (8) or (9) are calculatedBAnd SCIn key be C and element value that key is B.
(note: update is not assignment, is weighting)
(11)SAIn node whether by all secondary traversals, if so, (12) are thened follow the steps, if it is not, then jumping back to step
(6)。
(12) select with (2) in unduplicated destination node, execute step (2), until all nodes in network topology all
It is expressed study.
(13) representation method for passing through skip-gram term vector, by origin node A and set of context SARespectively as neural network
Output and input, the parameter learnt is the expression vector of node A.
(14) the expression vector of all nodes in network is obtained by step (13) traversal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811175451.3A CN109447261B (en) | 2018-10-09 | 2018-10-09 | Network representation learning method based on multi-order proximity similarity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811175451.3A CN109447261B (en) | 2018-10-09 | 2018-10-09 | Network representation learning method based on multi-order proximity similarity |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109447261A true CN109447261A (en) | 2019-03-08 |
| CN109447261B CN109447261B (en) | 2023-08-04 |
Family
ID=65546122
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811175451.3A Active CN109447261B (en) | 2018-10-09 | 2018-10-09 | Network representation learning method based on multi-order proximity similarity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109447261B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110348469A (en) * | 2019-05-21 | 2019-10-18 | 广东工业大学 | A kind of user's method for measuring similarity based on DeepWalk internet startup disk model |
| CN110417594A (en) * | 2019-07-29 | 2019-11-05 | 吉林大学 | Network establishing method, device, storage medium and electronic equipment |
| CN110826590A (en) * | 2019-09-20 | 2020-02-21 | 浙江工商大学 | Learner relationship strength measurement method and device integrating learning characteristics and learning network structural characteristics |
| CN111309922A (en) * | 2020-01-19 | 2020-06-19 | 清华大学 | Map construction method, accident classification method, device, computer equipment and medium |
| WO2020199745A1 (en) * | 2019-03-29 | 2020-10-08 | 创新先进技术有限公司 | Sample clustering method and device |
| CN113886688A (en) * | 2021-09-27 | 2022-01-04 | 鹏城实验室 | Object incidence relation prediction method and device, terminal device and storage medium |
| CN117811992A (en) * | 2024-02-29 | 2024-04-02 | 山东海量信息技术研究院 | Network bad information propagation inhibition method, device, equipment and storage medium |
| CN113886688B (en) * | 2021-09-27 | 2024-07-05 | 鹏城实验室 | Method, device, terminal equipment and storage medium for predicting association relation of objects |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020116A (en) * | 2012-11-13 | 2013-04-03 | 中国科学院自动化研究所 | Method for automatically screening influential users on social media networks |
| CN104021199A (en) * | 2014-06-16 | 2014-09-03 | 西安电子科技大学 | Function module detecting method based on node domination capacity similarity |
| CN104765825A (en) * | 2015-04-10 | 2015-07-08 | 清华大学 | Method and device for predicting social network links based on cooperative fusion theory |
| US20160004963A1 (en) * | 2013-02-22 | 2016-01-07 | Tokyo Institute Of Technology | Information processing apparatus, information processing method, and non-transitory computer readable medium |
-
2018
- 2018-10-09 CN CN201811175451.3A patent/CN109447261B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020116A (en) * | 2012-11-13 | 2013-04-03 | 中国科学院自动化研究所 | Method for automatically screening influential users on social media networks |
| US20160004963A1 (en) * | 2013-02-22 | 2016-01-07 | Tokyo Institute Of Technology | Information processing apparatus, information processing method, and non-transitory computer readable medium |
| CN104021199A (en) * | 2014-06-16 | 2014-09-03 | 西安电子科技大学 | Function module detecting method based on node domination capacity similarity |
| CN104765825A (en) * | 2015-04-10 | 2015-07-08 | 清华大学 | Method and device for predicting social network links based on cooperative fusion theory |
Non-Patent Citations (4)
| Title |
|---|
| ANDREW RODRIGUEZ等: "New multi-stage similarity measure for calculation of pairwise patent similarity in a patent citation network", 《SCIENTOMETRICS (2015)》 * |
| GUERRIERO等: "Power Law Distribution: Method of Multi-scale Inferential Statistics", 《JOURNAL OF MODEM MATHEMATICS FRONTIER (JMMF)》 * |
| 姜雅等: "基于节点相似度的网络社团检测算法研究", 《计算机科学》 * |
| 杜凌霞等: "概率图上的对象相似度计算", 《计算机研究与发展》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020199745A1 (en) * | 2019-03-29 | 2020-10-08 | 创新先进技术有限公司 | Sample clustering method and device |
| CN110348469A (en) * | 2019-05-21 | 2019-10-18 | 广东工业大学 | A kind of user's method for measuring similarity based on DeepWalk internet startup disk model |
| CN110417594A (en) * | 2019-07-29 | 2019-11-05 | 吉林大学 | Network establishing method, device, storage medium and electronic equipment |
| CN110417594B (en) * | 2019-07-29 | 2020-10-27 | 吉林大学 | Network construction method and device, storage medium and electronic equipment |
| CN110826590A (en) * | 2019-09-20 | 2020-02-21 | 浙江工商大学 | Learner relationship strength measurement method and device integrating learning characteristics and learning network structural characteristics |
| CN111309922A (en) * | 2020-01-19 | 2020-06-19 | 清华大学 | Map construction method, accident classification method, device, computer equipment and medium |
| CN111309922B (en) * | 2020-01-19 | 2023-11-17 | 清华大学 | Map construction method, accident classification device, computer equipment and medium |
| CN113886688A (en) * | 2021-09-27 | 2022-01-04 | 鹏城实验室 | Object incidence relation prediction method and device, terminal device and storage medium |
| CN113886688B (en) * | 2021-09-27 | 2024-07-05 | 鹏城实验室 | Method, device, terminal equipment and storage medium for predicting association relation of objects |
| CN117811992A (en) * | 2024-02-29 | 2024-04-02 | 山东海量信息技术研究院 | Network bad information propagation inhibition method, device, equipment and storage medium |
| CN117811992B (en) * | 2024-02-29 | 2024-05-28 | 山东海量信息技术研究院 | Network bad information propagation inhibition method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109447261B (en) | 2023-08-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109447261A (en) | A method of the network representation study based on multistage neighbouring similarity | |
| Ji et al. | Dual channel hypergraph collaborative filtering | |
| CN106649659B (en) | Social network-oriented link prediction system and method | |
| CN106709035B (en) | A kind of pretreatment system of electric power multidimensional panoramic view data | |
| CN103106279B (en) | Clustering method a kind of while based on nodal community and structural relationship similarity | |
| CN107193967A (en) | A kind of multi-source heterogeneous industry field big data handles full link solution | |
| CN102810113B (en) | A kind of mixed type clustering method for complex network | |
| CN104268271A (en) | Interest and network structure double-cohesion social network community discovering method | |
| CN107103000A (en) | It is a kind of based on correlation rule and the integrated recommended technology of Bayesian network | |
| CN106951524A (en) | Overlapping community discovery method based on node influence power | |
| CN107391542A (en) | A kind of open source software community expert recommendation method based on document knowledge collection of illustrative plates | |
| CN107273934A (en) | A kind of figure clustering method merged based on attribute | |
| Xiao et al. | Link prediction based on feature representation and fusion | |
| Wang et al. | A novel dual-graph convolutional network based web service classification framework | |
| Yigit et al. | Extended topology based recommendation system for unidirectional social networks | |
| CN116340646A (en) | Recommendation method for optimizing multi-element user representation based on hypergraph motif | |
| Chen et al. | Network dynamics in university-industry collaboration: A collaboration-knowledge dual-layer network perspective | |
| Chen et al. | Pre-training on dynamic graph neural networks | |
| Lu et al. | A unified link prediction framework for predicting arbitrary relations in heterogeneous academic networks | |
| CN103577899A (en) | Service composition method based on reliability prediction combined with QoS | |
| Bhatnagar et al. | Role of machine learning in sustainable engineering: a review | |
| Niu et al. | Semi-supervised plsa for document clustering | |
| Wang et al. | Emotion-based Independent Cascade model for information propagation in online social media | |
| CN105761152A (en) | Topic participation prediction method based on triadic group in social network | |
| Liu et al. | ICE: Information credibility evaluation on social media via representation learning |
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 |