CN107527295A - Dynamics community of Academic Teams based on tense coauthorship network finds method and its method for evaluating quality - Google Patents

Dynamics community of Academic Teams based on tense coauthorship network finds method and its method for evaluating quality Download PDF

Info

Publication number
CN107527295A
CN107527295A CN201710737012.6A CN201710737012A CN107527295A CN 107527295 A CN107527295 A CN 107527295A CN 201710737012 A CN201710737012 A CN 201710737012A CN 107527295 A CN107527295 A CN 107527295A
Authority
CN
China
Prior art keywords
mrow
node
msub
community
author
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
CN201710737012.6A
Other languages
Chinese (zh)
Other versions
CN107527295B (en
Inventor
黄芳
万文聪
王向前
张予琛
章成源
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.)
Central South University
Original Assignee
Central South 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 Central South University filed Critical Central South University
Priority to CN201710737012.6A priority Critical patent/CN107527295B/en
Publication of CN107527295A publication Critical patent/CN107527295A/en
Application granted granted Critical
Publication of CN107527295B publication Critical patent/CN107527295B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • 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

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The present invention proposes a kind of dynamics community of Academic Teams based on tense coauthorship network and finds algorithm, by establishing tense coauthorship network model, in detection coauthorship network in real time on the basis of author node importance, relation side Strength Changes, analysis and the evolution process for tracking Academic Teams in academic coauthorship network, community is created, extended, is shunk, is divided with withering away, to reach the purpose of dynamics community's discovery, algorithm accuracy is high.In addition, the present invention also proposes community of the Academic Teams division method for evaluating quality based on word feature to assess community's division quality, and experimental verification is carried out with public literature information record data set, test result indicates that the validity of the algorithm.

Description

Dynamics community of Academic Teams based on tense coauthorship network finds that method and its quality are commented Estimate method
Technical field
The present invention relates to a kind of dynamics community of Academic Teams based on tense coauthorship network to find method.
Background technology
State innovation system needs a collection of high-level research lab with the capability of independent innovation as its important composition portion Point.How effectively to observe with the differentiation of analysis and research colony, comprehensively and objectively assess Academic Teams overall performances, be science and technology The new problem selected with being faced in assessing of innovation colony.Research Team is the entirety of Long-term Collaboration self-assembling formation, and it is complicated Evolution process often lie in researcher cooperation publish an article, the various academic activities history notes such as joint study project In record, wherein, scientific and technical literature common data resource more fully have recorded the history rail that researcher cooperation is published an article Mark, therefore, academic coauthorship network is built using large-scale scientific and technological literature database, and find to manage with dynamics community on this basis The important subject that Academic Teams' evolution process is current Complex Networks Analysis field is excavated by with method.
Community structure in academic coauthorship network is the entirety that the long-term scientific cooperation of community member is formed, and generally has and compares Stable core member, an important factor for driving community develops are the transition of its core member.The dynamics community of early stage finds to calculate Method is to apply to the algorithm idea of static community discovery in dynamic network mostly, using two-step Taylor-Galerkin, i.e., first with static community Detection algorithm carries out community's division to network snapshots at different moments, then to entering between community's division at two neighboring moment Row matching.2007, Palla et al. [1] removed structure not using the node overlapping degree between CPM algorithms [2] detection community Evolvement between community in the same time.Hopcroft et al.2004 [3], Greene et al.2010 [4] utilize overlapping Degree similitude develops to community to be analyzed.Wang et al.2008 [5] are tracked using part important node to community. However, the static community division based on network snapshots can not describe the evolution Feature of community exactly.For this problem, Yang et al.2005 [6], Miller et al.2010 [7], Caravelli et al.2013 [8] etc. is moved using increment type State community discovery method, this kind of method obtains initial community first with state algorithm and divided, then according to network topology knot The change of structure changes to guide current time community to be divided on the basis of former community；This kind of method is easy to follow the trail of community Differentiation, but the network change for how defining each moment is this kind of method key issue to be solved.2010, Cazabet et al. [9] propose to regard as the change on side between nodes to be changed the interaction node with the time Become, and realize that dynamics community detects according to the inherent attribute of community and the interaction course of node.Rossetti et in 2016 Al. [10] propose TILES (Temporal Interactions a Local Edge Strategy) algorithm, in the algorithm Relation side can not only be added over time by considering network, and existing side may also disappear both factors pair Community influences caused by developing so that dynamics community finds that algorithm is more efficient, but does not account for importance and the friendship of node Mutual tightness degree is on influence caused by community structure.However, the relation meeting due to the status of Team Member and between them Change over time and change, this causes the importance in academic network of member node to change, so that Academic Teams Institutional framework produces differentiation, and the team structure that this differentiation result in using team core member as guiding develops；So In Academic Teams' community discovery, the change of character and the tightness degree of relation be influence that Academic Teams develop it is important because Element.
In addition, substantial amounts of document record is continuously added in bibliographic data base over time, Xue Zhehe Work relation will also produce change therewith, and this not only result in scholar's science network topology structure and changes, and be saved in network The importance of point also changes, and is to influence an important factor for academic community changes.
Therefore, it is necessary to sent out for a kind of new dynamics community of Academic Teams of the design of technical problem present in prior art Existing method.
The content of the invention
Technical problem solved by the invention is, in view of the shortcomings of the prior art, it is proposed that one kind collaborates net based on tense The dynamics community of Academic Teams of network finds algorithm (ATDD, Academic Team Dynamic Discovery), passes through detection Author node importance, relation side intensity (weight) and continuation change over time, and analyze with tracking academic coauthorship network middle school The evolution process of art team, community is created, extended, is shunk, is divided with withering away, to reach the mesh of dynamics community's discovery 's.In addition, the present invention also proposes community of the Academic Teams division method for evaluating quality based on word feature to assess community's division matter Amount, and carried out experimental verification with public literature information record data set, test result indicates that the validity of the algorithm.
Academic Teams member institute's role in institutional framework can be divided into team leader people, important member and general Member, correspond to core node, important node and slave node respectively in academic network, and be formed Academic Teams with this Community.Change over time, the academic relationship between Team Member can also change with the change of scientific cooperation behavior, learn Art relation or enhancing weaken or create or disappear, and this causes the importance in academic network of member node to change Become so that institute's role also produces change, this will cause Academic Teams' institutional framework produces to develop, and it is this develop it is past Past is that the behavior of Academic Teams core member plays a leading role, and result in the team structure using team core member as guiding and drills Become, this differentiation is also be reflected in the cooperation behavior that researcher constantly cooperates to deliver academic article.In public literature resource In database, documentation ＆ info data record will be constantly accumulated in database in the form of time series stream, result in formation of Author with time attribute collaborates the stream data of relation pair.The academic relationship network of researcher is collaborateing relation data Constantly there are side and node to be added in network under the driving of stream, important node and side are highlighted.Therefore the present invention is by right The when state property measurement of the importance of personage's node and relation weight assesses the nuclear of node, and use using core node as The iterative manner at center goes to find team.
Fig. 1 is that Academic Teams develop schematic diagram.When new data arrive, node and side in network can be updated, New core node may be produced, this will cause the formation of a new communities, as shown in Fig. 1 (a), add node 4 and side When (Isosorbide-5-Nitrae), node 1 becomes core node, forms new community.With the addition of relation pair, make the node outside community and be somebody's turn to do The contact of community becomes closely, as shown in Fig. 1 (c), adds node 5 and side (1,5), and make node 5 and community contacts change Obtain closely, we are added into community, and this increases the scale for making community.As shown in Fig. 1 (b), over time Some cooperative relationship can disappear, and cause node to be removed from community, and community no longer includes core node, then the community is dead.This The removal on the expired side of kind, can make some nodes and contacting for community become not close, it is necessary to which these nodes are removed from community, As shown in Fig. 1 (d), this diminishes the scale for making community.This addition operation will can be connected in a manner of iterating with community Node closely adds community, and therefore, the induced subgraph of community is a connected graph, as shown in Fig. 1 (e), in Liang Ge societies Edged between area, one of community's major part node is caused all to add another community, then the two communities are merged into one Community.The side removed in community may destroy this connectedness, core be present if removing side and causing the induced subgraph of community to contain The connected component of heart node, then the community is split into multiple communities, as shown in Fig. 1 (f), community's internal edges disappear, and split into Liang Ge communities.
Based on principles above, technical scheme provided by the present invention is：
A kind of dynamics community of Academic Teams based on tense coauthorship network finds method, by establishing tense coauthorship network mould Type, on the basis of detection coauthorship network interior joint importance in real time and relationship strength change, created, extended, received with community Contracting, division go to realize that dynamics community divides with extinction strategy；Specifically include following steps：
Step 1, the coauthorship network G for defining tt=(Vt,Et), whereinFor t Coauthorship network in author node set,Represent author node viIn the importance of t,Over time Passage, the importance that author often delivers an article then its node add 1, and author's more more important property of publishing an article is higher；Be t coauthorship network in side collection；Author node viAnd vjClose Work delivered article, then had side e between themi,jIt is connected；Represent author node υiAnd υjBetween side ei,jIn t Weight,The bigger explanation υ of weightiAnd υjBetween relation it is closer, υiAnd υjOften collaborate and deliver an article,'s Value is then once updated, shown in the weight calculation such as formula (1) on side：
Wherein, n represents υiAnd υjCollaborate author's quantity in the article delivered；Formula (1) represents υiAnd υjOften collaborate and deliver one Piece article, corresponding side ei,jWeight side number of the increment between all authors of this article inverse, show H3 segment Its more weight increment are smaller；
For side ei,jIn the residual life of t,υiAnd υjOften collaborate and deliver an article, then assignmentFrom υiAnd υjLast time collaboration is calculated from publishing an article, if υiAnd υjAlready exceed T1Time does not have collaboration to deliver text Chapter, thenIf υiAnd υjAlready exceed T2Time does not have collaboration to publish an article, thenIfV beforeiAnd vjAgain Secondary collaboration is published an article, then is resetWherein T1And T2For time threshold, rule of thumb value；
Step 2, initialization coauthorship network G0=(V0,E0),V0=Φ, E0=Φ, C0=Φ；
Step 3, input documentation ＆ info record R={ r1,r2,...,rt... }, wherein rtIt is one contain title, make The article that person, keyword and article deliver the information such as time delivers record, rtIn R time-sequencing, t r are delivered by articlet Sequence number in R, a corresponding moment；
Step 4, each article one by one in R deliver record rtUpdate coauthorship network Gt, carry out dynamics community and draw Point, obtain Ct={ c0,c1,...,ck,...}：
Step 4.1, make t=1；
Step 4.2, one article of taking-up delivers record r from Rt；By rtIn all authors formed author between collaborate relation Node pair, and be added in coauthorship network, the importance of each node and the weight on each bar side in coauthorship network are then updated, And according to rtUpdate the residual life on side：
If rtIn include author node vi, then author node v is updatediIt is in the importance of tIf rt In include author node viAnd υj, and author node viAnd vjBetween be connected originally without side, then in author node viAnd vjBetween Increase a line ei,j, then update side ei,jIt is in the weight of tN represents rtIn include Author node quantity；And update side ei,jIt is in the residual life of tIf rtIn include author node viAnd vj, and make Person's node viAnd υjBetween had side ei,jIt is connected, then directly updates side e as stated abovei,jIn the weight and residue of t Life-span；If rtAuthor node υ is included when middle differentiAnd υj, and author node viAnd vjBetween had side ei,jIt is connected, then basis Documentation ＆ info record judges t distance viAnd vjLast time collaborates the length for the time published an article, if being less than T1, thenIf it is more than or equal to T1And it is less than T2, thenIf it is more than or equal to T2, thenWherein T1And T2For time threshold, root According to experience value；
When document record is added step-wise to document databse with time series, the importance and work of author node in coauthorship network The tight ness rating of cooperative relationship between person will also be changed, and the topological structure for causing community of Academic Teams is also changed, this Kind change can cause the differentiation of community structure, including create, extend, dividing with withering away.
Step 4.3, according to rtIn article deliver the time, expired author's cooperative relationship in current coauthorship network is moved Remove, investigated further according to Academic Teams' community's criteria for classifying due to removing side to influence caused by community, perform corresponding community and receive Contracting, division and extinction；
Step 4.4, the increase due to side investigated to influence caused by community according to Academic Teams community's criteria for classifying, performed Corresponding community creates and extension；
Step 4.5, so far, community of the Academic Teams division in t coauthorship network is obtained, is expressed as Ct={ c0, c1,...,ck,...}；
Step 4.6, t=t+1 is made, repeat step 4.2~4.5, record r is delivered until taking out the last item article from Rt And untill performing aforesaid operations, obtain final Academic Teams' community division result；
In the step 4.3 and 4.4, coauthorship network G is definedtIn important node, Qiang Bian, slave node and core node For：
1) important node：When the importance of a node is not less than the average importance of all non-orphaned nodes in network When, this node is an important node；Wherein non-orphaned node is that have the node that side is connected with other nodes；
2) strong side：When a connection two important nodes while weight be not less than all in network while average weight When, this edge is exactly strong side；
3) slave node and core node：If the two node v connected by strong sideiAnd vjImportant sexual satisfaction：
And the weight on the side of two nodes meets：
Then vjIt is viSlave node；
Wherein, M is to form the number that an Academic Teams at least need, and rule of thumb value, typically takes M=4；X is represented With vjAll neighbor node v connected by strong sidex,It is with vjFor the weight sum on all strong sides of end points；
Formula (2) and formula (3) show slave node vjImportance be far smaller than node viImportance, and viAnd vjIt Between the weight on side be greater than with vjFor the intermediate value of the weight sum on all strong sides of end points；
When a node has M-1 slave node, then it is a core node；
Academic Teams' community's criteria for classifying refers to：
The group node connected by strong side is formed into a community of Academic Teams, must be wrapped in a community of Academic Teams Formula (4) must is fulfilled for containing all non-core nodes in core node and this community, i.e., is contacted with community's interior nodes close Degree is more than the tight ness rating contacted with community's exterior node；
Wherein, viFor the node in community,Represent in community with viFor the weight sum on all strong sides of end points,Represent in whole coauthorship network with viFor the weight sum on all strong sides of end points.
Further, the step 4.3 specifically includes following steps：
For every a line in t coauthorship network, ifThen the side is removed from network；After removing side Three kinds of situations are divided to carry out corresponding community's contraction, division and extinctions processing：
(a) shrink：Author node vuAnd vvBelong to community ck, and ckInduced subgraph be a connected graph, and core be present Heart node, now, if vuAnd vvFormula (4) is unsatisfactory for, then is removed them from community；Any one node is from community After removal, its neighbor node and community in community is completely embedded degree and can also changed, thus must iteration judge Whether the neighbor node for removing node meets formula (4), continues to remove it if being unsatisfactory for, until the nodes in community (scale of community) is less than M, or no node will remove；
(b) divide:If author node vuAnd vvBelong to community ck, and ckInduced subgraph contain multiple connected subgraphs, And the connected subgraph having includes core node, some connected subgraphs are free of core node, then by the connection comprising core node Subgraph forms new communities' (each connected subgraph for including core node splits into a new communities), not comprising core node Connected subgraph in node all from CkMiddle removal；
(c) wither away：If author node vuAnd vvBelong to community ck, and ckInduced subgraph contain multiple connected subgraphs, And each connected subgraph does not include core node, then community CkWither away；
Further, in the step 4.4, community creates and extended operation is：
Create：To rtIn any group of author composition node to (u, v), if author node vuAnd vvIt is not belonging to any Community, it is being not belonging to two node v of any communityuAnd vvBetween edged, make vuAs vvSlave node, and vvFor core Node, then one is created with vvFor the new communities of core node, the adjacent node for meeting formula (4) is iteratively added to this Community；
Extension：If increase by two node v on sideuAnd vvIn, if node vuBelong to community ckAnd vuNeighbor node vvIt is full Sufficient formula (4), then by vvAdd vuAffiliated community ckIn, and iteration judges whether the neighbor node of newly added node meets public affairs Formula (4), community is also added if meeting；Such community's expansion can cause another community's atrophy, substantially generate community Merging.
Further, in the step 4.2, T11 year is arranged to, T2It is arranged to 2 years.
Present invention also offers a kind of community of Academic Teams based on word feature to divide method for evaluating quality；
In general, Academic Teams are made up of the researcher with joint research interest, the division of community of Academic Teams Quality can be reflected with the similitude of community's interior nodes.The keyword of article reflects author in documentation ＆ info record Research interest, can thus form the research interest keyword set of author, the research interest similarities of Academic Teams can be with Measured with the similarity of the keyword of community member.But because academic article keyword has high discrimination, use Word feature goes its validity of measurement team's research interest similarity poor, so, the present invention proposes the academic group based on word feature Community of team division method for evaluating quality.The word of the keyword for the article delivered with author is characterized as that each author establishes one Word vector.
Key_charater(vi)={ ch1:Count1, ch2:Count2 ..., chn:countn} (5)
Formula (5) is represented in author node viThe keyword set published an article is made up of n keyword, wherein keyword chn The number that appearance is accumulated in the keyword of all articles delivered of the author is countn.Interest Similarity between author is used They share proportion of the counting of keyword shared by their keyword tale to measure, calculation formula such as (6) institute Show
Wherein,WithRepresent keyword ch in author node v respectivelyiAnd vjGo out in the article delivered Existing number；
Community ckInterest degree of correlation correlation (ck) Interest Similarity between all authors in community is averaged Value, calculation formula as shown in (7)
Wherein, | ck| represent community ckIn author's quantity；
Community division quality PartitionQuality (Ct) for all communities the interest degree of correlation weighted sum, such as formula (8) shown in, the weight of each community is author's quantity total in author's quantity in the community divided by all communities.
Beneficial effect：
Dynamics community proposed by the present invention finds that algorithm is used for finding Academic Teams, and algorithm uses sequential stream data conduct Input, and the tight ness rating for being interacted between Team Member's importance and member changes with time come to team's community structure Caused to influence to realize that dynamics community finds, it need not be matched between community's division at different moments, Neng Gougen The evolutionary process of community is continuously found according to time series.Further, since introduce tense node importance and relation side when State weight, eliminate insignificant node community is divided caused by influence, according to the different residual life of each edge, more rationally Ground describes the characteristics of relation retention time difference between different members so that community discovery algorithm is more accurate.It is in addition, of the invention The division method for evaluating quality of the community of Academic Teams based on word feature is also proposed to assess community's division quality, and with public text Offer information record data set and carried out experimental verification, test result indicates that the validity of the algorithm.
Brief description of the drawings
Fig. 1 is community of Academic Teams evolution schematic diagram；Fig. 1 (a)~(f) is the establishment of community, extinction, extension, receipts respectively Contracting, merge and divide；
Fig. 2 is accumulation node of graph and side variation tendency；Fig. 2 (a)~(d) is the author node number of four data sets respectively And there is the number on collaboration relation side between them with the accumulation situation of time；
Fig. 3 is that ATDD and TILES algorithms community division quality changes with time；Fig. 3 (a)~(d) is four numbers respectively Changed with time according to community's division quality of collection；
Fig. 4 is that ATDD and TILES algorithms community quantity changes with time；Fig. 4 (a)~(d) is four data sets respectively Community's quantity change with time.
Embodiment
The present invention is described in more detail below in conjunction with the drawings and specific embodiments.
The present invention proposes a kind of dynamics community of Academic Teams based on tense coauthorship network and finds algorithm, passes through when establishing State coauthorship network model, in detection coauthorship network in real time on the basis of author node importance, relation side Strength Changes, point Analysis and the evolution process for tracking Academic Teams in academic coauthorship network, are created to community, are extended, are shunk, divided and are disappeared Die, to reach the purpose of dynamics community's discovery, algorithm accuracy is high.In addition, the present invention also proposes the academic group based on word feature Community of team division method for evaluating quality divides quality to assess community, is carried out below with public literature information record data set real Checking, to prove the validity of the algorithm.
Documentation ＆ info record in disclosed in National IP Network's document resource database is used as experimental data, the time of data For span from 2000 to 2016 year, every document record includes article name, author, mechanism, keyword, summary and publication time Input traffic is formed etc. information, and by publication time, goes the validity of verification algorithm.The system environments of experiment is PC CPU i5-3337U, 8G RAM, Windows10, python3.5, networkx1.1.
1 experimental data
By documentation ＆ info record by mechanism categories into 4 data sets, each data set is respectively data set (a) 5925, number According to 4586 collection (b) 5892, data set (c) 3792 and data set (d) records, four subgraphs are this four data respectively in Fig. 2 The author node number of collection and there is the number on collaboration relation side between them with the accumulation situation of time.
2 contrast experiments
For TILES (the Temporal Interactions a Local Edge Strategy) sides with document [10] Method is contrasted, and we are 2 years to the life-span that each edge is fixed, and four data sets are performed with ATDD algorithms respectively and TILES is calculated Method, and assessed with Academic Teams' interest-degree community method for evaluating quality based on word feature, its result is as shown in Figure 3.By In TILES be the community discovery based on triangular structure, so, when data volume is less, TILES community division quality Higher than ATDD.Passage data volume constantly increases again over time, and community's division quality of ATDD algorithms exceedes soon TILES, and keep higher division quality.
Table 1 is contrast of the two kinds of algorithms of ATDD and TILES for community's division quality of four groups of test datas, is distinguished in table The average value that each data set divides in having time community of institute, and the overall average of four data sets are listed, ATDD is each Community's division quality on data set all performs better than TILES.
Table 1.ATDD and TILES algorithms community divides quality versus
ATDD and TILES algorithms community division numbers change with time as can be seen from Figure 4, with nodes and side Several accumulations, the number of four groups of community of data team divisions of ATDD algorithms are relatively stable；Community's division number of TILES algorithms Rise with the increase of node and the number on side, when network size is very big, community's division number increases severely.Table 2 is ATDD and TILES algorithms community division number contrast, wherein, community's division number of each data set is the average value at all moment, is shown So, community's number that TILES has found is far longer than ATDD, because data set is set by research institution, corresponding to each mechanism Research team's number be limited, so, the communities of ATDD algorithms division number is more reasonable.
Table 2.ATDD and TILES algorithms community division number contrast
Community's number Data set (a) Data set (b) Data set (c) Data set (d) Overall average
ATDD 11 15 13 11 12
TILES 144 152 230 119 161
Bibliography
1.Palla G,Barabási A L,Vicsek T.Quantifying social group evolution. [J].Nature,2007,
446(7136):664-667.
2.Palla G,Derényi I,Farkas I,et al.Uncovering the overlapping community structure of complex networks in nature and society[J].Nature,2005, 435(7043):814-818.
3.Hopcroft J,Khan O,Kulis B,et al.Tracking evolving communities in large linked networks[J].Proceedings of the National Academy of Sciences, 2004,101(suppl 1):5249-5253.
4.Greene D,Doyle D,Cunningham P.Tracking the Evolution of Communities in Dynamic Social Networks[C]//International Conference on Advances in Social Networks Analysis and Mining,Asonam 2010,Odense,Denmark,August.2010:176-183.
5.Wang Y,Wu B,Du N.Community Evolution of Social Network:Feature, Algorithm and Model[J].arXiv:0804.4356v1[physics.soc-ph],28Apr 2008.
6.Yang B,Liu D Y.Incremental algorithm for detecting community structure in dynamic networks[C]//Machine Learning and Cybernetics, 2005.Proceedings of 2005International Conference on.IEEE,2005,4:2284-2290.
7.Miller K,Eliassi-Rad T.Continuous time group discovery in dynamic graphs[R]. Lawrence Livermore National Laboratory(LLNL),Livermore,CA,2010.
8.Caravelli P,Wei Y,Subak D,et al.Understanding evolving group structures in time-varying networks[C]//Proceedings of the 2013IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining.ACM,2013:142-148.
9.Cazabet R,Amblard F,Hanachi C.Detection of overlapping communities in dynamical social networks[C]//Social Computing(SocialCom),2010 IEEE Second International Conference on.IEEE,2010:309-314.
10.Rossetti G,Pappalardo L,Pedreschi D,et al.Tiles:an online algorithm for community discovery in dynamic social networks[J].Machine Learning,2016:1-29.

Claims (5)

1. a kind of dynamics community of Academic Teams based on tense coauthorship network finds method, it is characterised in that comprises the following steps：
Step 1, the coauthorship network G for defining tt=(Vt,Et), whereinFor t Coauthorship network in author node set,Represent author node viIn the importance of t, Be t coauthorship network in side collection, author node viAnd vjClose Work delivered article, then had side e between themi,jIt is connected；Represent author node viAnd vjBetween side ei,jIn the power of t Weight, For side ei,jIn the residual life of t,viAnd vjOften collaborate and deliver an article, then assignmentFrom viAnd vjLast time collaboration is calculated from publishing an article, if viAnd vjAlready exceed T1Time does not have collaboration to deliver text Chapter, thenIf viAnd vjAlready exceed T2Time does not have collaboration to publish an article, thenIfV beforeiAnd vjAgain Collaboration is published an article, then is resetWherein T1And T2For time threshold, rule of thumb value；
Step 2, initialization coauthorship network G0=(V0,E0),V0=Φ, E0=Φ, C0=Φ；
Step 3, input documentation ＆ info record R={ r1,r2,...,rt... }, wherein rtIt is one and contains title, author, pass The article that key word and article deliver the time delivers record, rtIn R time-sequencing, t r are delivered by articletSequence number in R is right Answer a moment；
Step 4, each article one by one in R deliver record rtUpdate coauthorship network Gt, dynamics community's division is carried out, is obtained To Ct={ c0,c1,...,ck,...}：
Step 4.1, make t=1；
Step 4.2, one article of taking-up delivers record r from Rt, by rtIn all authors formed author between collaborate relation node It is right, and be added in coauthorship network, then update importance, the weight on each bar side of each node in coauthorship network, and according to rtUpdate the residual life on side：
If rtIn include author node vi, then author node v is updatediIt is in the importance of tIf rtMiddle bag V containing author nodeiAnd vj, and author node viAnd vjBetween be connected originally without side, then in author node viAnd vjBetween increase A line ei,j, then update side ei,jIt is in the weight of t：
<mrow> <msubsup> <mi>w</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mi>t</mi> </msubsup> <mo>=</mo> <msubsup> <mi>w</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mrow> <mi>t</mi> <mo>-</mo> <mn>1</mn> </mrow> </msubsup> <mo>+</mo> <mfrac> <mn>2</mn> <mrow> <mi>n</mi> <mo>*</mo> <mrow> <mo>(</mo> <mi>n</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>
Wherein, n represents rtIn the author node quantity that includes, and update side ei,jIt is in the residual life of tIf rtIn Include author node viAnd vj, and author node viAnd vjBetween had side ei,jIt is connected, then directly updates side as stated above ei,jIn the weight and residual life of t；If rtAuthor node v is included when middle differentiAnd vj, and author node viAnd vjBetween There is side ei,jIt is connected, then t distance v is judged according to documentation ＆ info recordiAnd vjLast time collaborates the time published an article Length, if being less than T1, thenIf it is more than or equal to T1And it is less than T2, thenIf it is more than or equal to T2, then
Step 4.3, according to rtIn article deliver the time, expired author's cooperative relationship in current coauthorship network is removed, then Investigated according to Academic Teams' community's criteria for classifying due to removing side to influence caused by community, perform corresponding community and shrink, divide Split and wither away；
Step 4.4, the increase due to side investigated to influence caused by community according to Academic Teams community's criteria for classifying, performed corresponding Community create and extension；
Step 4.5, so far, community of the Academic Teams division in t coauthorship network is obtained, is expressed as Ct={ c0,c1,..., ck,...}；
Step 4.6, t=t+1 is made, repeat step 4.2~4.5, record r is delivered until taking out the last item article from RtPerform Untill aforesaid operations, final Academic Teams' community division result is obtained；
In the step 4.3 and 4.4, coauthorship network G is definedtIn important node, Qiang Bian, slave node and core node be：
1) important node：When the importance of a node is not less than the average importance of all non-orphaned nodes in network, this Individual node is an important node；Wherein non-orphaned node is that have the node that side is connected with other nodes；
2) strong side：When a connection two important nodes while weight be not less than all in network while average weight when, this Bar is when being exactly strong；
3) slave node and core node：If the two node v connected by strong sideiAnd vjImportant sexual satisfaction：
<mrow> <msubsup> <mi>p</mi> <mi>i</mi> <mi>t</mi> </msubsup> <mo>&amp;GreaterEqual;</mo> <mrow> <mo>(</mo> <mi>M</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>&amp;CenterDot;</mo> <msubsup> <mi>p</mi> <mi>j</mi> <mi>t</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow>
And the weight on the side of two nodes meets：
<mrow> <msubsup> <mi>w</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mi>t</mi> </msubsup> <mo>&amp;GreaterEqual;</mo> <mn>0.5</mn> <mo>&amp;CenterDot;</mo> <msubsup> <mi>&amp;Sigma;w</mi> <mrow> <mi>j</mi> <mo>,</mo> <mi>x</mi> </mrow> <mi>t</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow>
Then vjIt is viSlave node；
Wherein, M is to form the number that an Academic Teams at least need, rule of thumb value；X is represented and vjConnected by strong side All neighbor node vx,It is with vjFor the weight sum on all strong sides of end points；
When a node has M-1 slave node, then it is a core node；
Academic Teams' community's criteria for classifying refers to：
The group node connected by strong side is formed into a community of Academic Teams, core must be included in a community of Academic Teams All non-core nodes must are fulfilled for formula (4) in heart node and this community, i.e., the tight ness rating contacted with community's interior nodes is big In the tight ness rating contacted with community's exterior node；
<mrow> <msubsup> <mi>&amp;Sigma;w</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>x</mi> </mrow> <mi>t</mi> </msubsup> <mo>&amp;GreaterEqual;</mo> <mn>0.5</mn> <mo>&amp;CenterDot;</mo> <msubsup> <mi>&amp;Sigma;w</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>y</mi> </mrow> <mi>t</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow>
Wherein, viFor the node in community,Represent in community with viFor the weight sum on all strong sides of end points, Represent in whole coauthorship network with viFor the weight sum on all strong sides of end points.
2. the Academic Teams dynamics community according to claim 1 based on tense coauthorship network finds method, its feature exists In the step 4.3 specifically includes following steps：For every a line in t coauthorship network, ifThen should While removed from network；Point three kinds of situations carry out corresponding community's contraction, division and extinction processing after removing side：
(a) shrink：Author node vuAnd vvBelong to community ck, and ckInduced subgraph be a connected graph, and core section be present Point, now, if vuAnd vvFormula (4) is unsatisfactory for, then is removed them from community, and iteration judges to remove the neighbours of node Whether node meets formula (4), continues to remove it if being unsatisfactory for, and until the nodes in community are less than M, or does not save Point will remove；
(b) divide:If author node vuAnd vvBelong to community ck, and ckInduced subgraph contain multiple connected subgraphs, and have Connected subgraph include core node, some connected subgraphs are free of core node, then by the connected subgraph comprising core node New communities are formed, the node in the connected subgraph not comprising core node is all from CkMiddle removal；
(c) wither away：If author node vuAnd vvBelong to community ck, and ckInduced subgraph contain multiple connected subgraphs, and often Individual connected subgraph does not all include core node, then community CkWither away；
3. the Academic Teams dynamics community according to claim 1 based on tense coauthorship network finds method, its feature exists In in the step 4.4, community creates and extended operation is：
Create：To rtIn any group of author composition node to (u, v), if author node vuAnd vvAny community is not belonging to, It is being not belonging to two node v of any communityuAnd vvBetween edged, make vuAs vvSlave node, and vvFor core node, then One is created with vvFor the new communities of core node, the adjacent node for meeting formula (4) is iteratively added to the community；
Extension：If increase by two node v on sideuAnd vvIn, if node vuBelong to community ckAnd vuNeighbor node vvMeet public Formula (4), then by vvAdd vuAffiliated community ckIn, and iteration judges whether the neighbor node of newly added node meets formula (4), community is also added if meeting.
4. the Academic Teams dynamics community according to claim 1 based on tense coauthorship network finds method, its feature exists In, in the step 4.2, T11 year is arranged to, T2It is arranged to 2 years.
5. a kind of community of Academic Teams division method for evaluating quality based on word feature, it is characterised in that for claim 1 The Academic Teams community division result C that method any one of~3 obtainst={ c0,c1,...,ck... } and quality enter Row is assessed：
First, the word of the keyword for the article delivered with author is characterized as that each author establishes a word vector：
Key_charater(vi)={ ch1:count1,ch2:count2,…,chn:countn} (5)
Formula (5) is represented in author node viThe keyword set published an article is made up of n keyword, wherein keyword chnIn the work The number that appearance is accumulated in the keyword of all articles delivered of person is countn；
Then, author is measured with the counting that keyword is shared between author proportion shared in their keyword tale Between Interest Similarity, calculation formula as shown in (6)
<mrow> <mi>S</mi> <mi>i</mi> <mi>m</mi> <mi>i</mi> <mi>l</mi> <mi>a</mi> <mi>r</mi> <mi>i</mi> <mi>t</mi> <mi>y</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msub> <mi>&amp;Sigma;</mi> <mrow> <mi>c</mi> <mi>h</mi> <mo>&amp;Element;</mo> <mi>K</mi> <mi>e</mi> <mi>y</mi> <mo>_</mo> <mi>c</mi> <mi>h</mi> <mi>a</mi> <mi>r</mi> <mi>a</mi> <mi>t</mi> <mi>e</mi> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;cap;</mo> <mi>K</mi> <mi>e</mi> <mi>y</mi> <mo>_</mo> <mi>c</mi> <mi>h</mi> <mi>a</mi> <mi>r</mi> <mi>a</mi> <mi>t</mi> <mi>e</mi> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> </msub> <msub> <mi>count</mi> <msub> <mi>v</mi> <mi>i</mi> </msub> </msub> <mrow> <mo>(</mo> <mi>c</mi> <mi>h</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>count</mi> <msub> <mi>v</mi> <mi>j</mi> </msub> </msub> <mrow> <mo>(</mo> <mi>c</mi> <mi>h</mi> <mo>)</mo> </mrow> </mrow> <mrow> <msub> <mi>&amp;Sigma;</mi> <mrow> <mi>c</mi> <mi>h</mi> <mo>&amp;Element;</mo> <mi>K</mi> <mi>e</mi> <mi>y</mi> <mo>_</mo> <mi>c</mi> <mi>h</mi> <mi>a</mi> <mi>r</mi> <mi>a</mi> <mi>t</mi> <mi>e</mi> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;cup;</mo> <mi>K</mi> <mi>e</mi> <mi>y</mi> <mo>_</mo> <mi>c</mi> <mi>h</mi> <mi>a</mi> <mi>r</mi> <mi>a</mi> <mi>t</mi> <mi>e</mi> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> </msub> <msub> <mi>count</mi> <msub> <mi>v</mi> <mi>i</mi> </msub> </msub> <mrow> <mo>(</mo> <mi>c</mi> <mi>h</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>count</mi> <msub> <mi>v</mi> <mi>j</mi> </msub> </msub> <mrow> <mo>(</mo> <mi>c</mi> <mi>h</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow>
Wherein, Similarity (vi,vj) represent author node viAnd vjInterest Similarity；countviAnd count (ch)vj (ch) represent keyword ch in author node v respectivelyiAnd vjThe number occurred in the article delivered；
Community c is calculated againkThe average value of Interest Similarity between interior all authors is as community ckThe interest degree of correlation correlation(ck), calculation formula as shown in (7)
<mrow> <mi>c</mi> <mi>o</mi> <mi>r</mi> <mi>r</mi> <mi>e</mi> <mi>l</mi> <mi>a</mi> <mi>t</mi> <mi>i</mi> <mi>o</mi> <mi>n</mi> <mrow> <mo>(</mo> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <mn>2</mn> <mo>&amp;CenterDot;</mo> <msub> <mo>&amp;Sigma;</mo> <mrow> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> <mo>&amp;Element;</mo> <msub> <mi>c</mi> <mi>k</mi> </msub> </mrow> </msub> <mi>S</mi> <mi>i</mi> <mi>m</mi> <mi>i</mi> <mi>l</mi> <mi>a</mi> <mi>r</mi> <mi>i</mi> <mi>t</mi> <mi>y</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>|</mo> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>|</mo> <mo>&amp;CenterDot;</mo> <mrow> <mo>(</mo> <mo>|</mo> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>|</mo> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow>
Wherein, | ck| represent community ckIn author's quantity；
Finally, the weighted sum of the interest degree of correlation of all communities is calculated as community division quality PartitionQuality (Ct), The weight of wherein each community is total author's quantity in author's quantity divided by all communities in the community, calculation formula such as formula (8) shown in：
<mrow> <mi>P</mi> <mi>a</mi> <mi>r</mi> <mi>t</mi> <mi>i</mi> <mi>t</mi> <mi>i</mi> <mi>o</mi> <mi>n</mi> <mi>Q</mi> <mi>u</mi> <mi>a</mi> <mi>l</mi> <mi>i</mi> <mi>t</mi> <mi>y</mi> <mrow> <mo>(</mo> <msub> <mi>C</mi> <mi>t</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mo>&amp;Sigma;</mo> <mrow> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>&amp;Element;</mo> <msub> <mi>C</mi> <mi>t</mi> </msub> </mrow> </munder> <mfrac> <mrow> <mo>|</mo> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>|</mo> </mrow> <mrow> <msub> <mo>&amp;Sigma;</mo> <mrow> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>&amp;Element;</mo> <msub> <mi>C</mi> <mi>t</mi> </msub> </mrow> </msub> <mo>|</mo> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>|</mo> </mrow> </mfrac> <mo>*</mo> <mi>c</mi> <mi>o</mi> <mi>r</mi> <mi>r</mi> <mi>e</mi> <mi>l</mi> <mi>a</mi> <mi>t</mi> <mi>i</mi> <mi>o</mi> <mi>n</mi> <mrow> <mo>(</mo> <msub> <mi>c</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>8</mn> <mo>)</mo> </mrow> <mo>.</mo> </mrow>
CN201710737012.6A 2017-08-24 2017-08-24 Academic team dynamic community discovery method based on temporal co-occurrence network and quality evaluation method thereof Active CN107527295B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710737012.6A CN107527295B (en) 2017-08-24 2017-08-24 Academic team dynamic community discovery method based on temporal co-occurrence network and quality evaluation method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710737012.6A CN107527295B (en) 2017-08-24 2017-08-24 Academic team dynamic community discovery method based on temporal co-occurrence network and quality evaluation method thereof

Publications (2)

Publication Number Publication Date
CN107527295A true CN107527295A (en) 2017-12-29
CN107527295B CN107527295B (en) 2021-04-30

Family

ID=60682220

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710737012.6A Active CN107527295B (en) 2017-08-24 2017-08-24 Academic team dynamic community discovery method based on temporal co-occurrence network and quality evaluation method thereof

Country Status (1)

Country Link
CN (1) CN107527295B (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110070364A (en) * 2019-03-27 2019-07-30 北京三快在线科技有限公司 Method and apparatus, storage medium based on the fraud of graph model detection clique
CN110110074A (en) * 2019-05-10 2019-08-09 齐鲁工业大学 A kind of timing data in literature analysis method and device based on Dynamic Network Analysis
CN110334264A (en) * 2019-06-27 2019-10-15 北京邮电大学 A kind of community detection method and device for isomery dynamic information network
CN111047453A (en) * 2019-12-04 2020-04-21 兰州交通大学 Detection method and device for decomposing large-scale social network community based on high-order tensor
CN111428056A (en) * 2020-04-26 2020-07-17 中国烟草总公司郑州烟草研究院 Method and device for constructing scientific research personnel cooperative community
CN112100452A (en) * 2020-09-17 2020-12-18 京东数字科技控股股份有限公司 Data processing method, device, equipment and computer readable storage medium
CN112463977A (en) * 2020-10-22 2021-03-09 三盟科技股份有限公司 Community mining method, system, computer and storage medium based on knowledge graph
WO2022198947A1 (en) * 2021-03-24 2022-09-29 南方科技大学 Method and apparatus for identifying close-contact group, and electronic device and storage medium

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101075942A (en) * 2007-06-22 2007-11-21 清华大学 Method and system for processing social network expert information based on expert value progation algorithm
CN103744846A (en) * 2013-08-13 2014-04-23 北京航空航天大学 Multidimensional dynamic local knowledge map and constructing method thereof

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101075942A (en) * 2007-06-22 2007-11-21 清华大学 Method and system for processing social network expert information based on expert value progation algorithm
CN103744846A (en) * 2013-08-13 2014-04-23 北京航空航天大学 Multidimensional dynamic local knowledge map and constructing method thereof

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
邹志科: "时态关联规则挖掘算法研究及其在学术合作关系挖掘中的应用", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110070364A (en) * 2019-03-27 2019-07-30 北京三快在线科技有限公司 Method and apparatus, storage medium based on the fraud of graph model detection clique
CN110110074A (en) * 2019-05-10 2019-08-09 齐鲁工业大学 A kind of timing data in literature analysis method and device based on Dynamic Network Analysis
CN110334264A (en) * 2019-06-27 2019-10-15 北京邮电大学 A kind of community detection method and device for isomery dynamic information network
CN111047453A (en) * 2019-12-04 2020-04-21 兰州交通大学 Detection method and device for decomposing large-scale social network community based on high-order tensor
CN111428056A (en) * 2020-04-26 2020-07-17 中国烟草总公司郑州烟草研究院 Method and device for constructing scientific research personnel cooperative community
CN112100452A (en) * 2020-09-17 2020-12-18 京东数字科技控股股份有限公司 Data processing method, device, equipment and computer readable storage medium
CN112100452B (en) * 2020-09-17 2024-02-06 京东科技控股股份有限公司 Method, apparatus, device and computer readable storage medium for data processing
CN112463977A (en) * 2020-10-22 2021-03-09 三盟科技股份有限公司 Community mining method, system, computer and storage medium based on knowledge graph
WO2022198947A1 (en) * 2021-03-24 2022-09-29 南方科技大学 Method and apparatus for identifying close-contact group, and electronic device and storage medium

Also Published As

Publication number Publication date
CN107527295B (en) 2021-04-30

Similar Documents

Publication Publication Date Title
CN107527295A (en) Dynamics community of Academic Teams based on tense coauthorship network finds method and its method for evaluating quality
Berahmand et al. A preference random walk algorithm for link prediction through mutual influence nodes in complex networks
CN104102745B (en) Complex network community method for digging based on Local Minimum side
Kossinets Effects of missing data in social networks
Marcus et al. Rage–a rapid graphlet enumerator for large networks
Bansal et al. Fast community detection for dynamic complex networks
Ilhan et al. Feature identification for predicting community evolution in dynamic social networks
CN104636426A (en) Multi-factor comprehensive quantitative analysis and sorting method for academic influences of scientific research institutions
Tajeuna et al. Tracking the evolution of community structures in time-evolving social networks
Liu et al. The degree-related clustering coefficient and its application to link prediction
Chen et al. Building and analyzing a global co-authorship network using google scholar data
Onel et al. The structure and analysis of nanotechnology co-author and citation networks
Šubelj et al. Convexity in scientific collaboration networks
CN103678279B (en) Personage&#39;s uniqueness recognition methods based on heterogeneous network temporal meaning similarity of paths
Li et al. Evolutionary community discovery in dynamic social networks via resistance distance
Adhikari et al. Propagation-based temporal network summarization
Zhang et al. Scientific relatedness and intellectual base: a citation analysis of un-cited and highly-cited papers in the solar energy field
Yang et al. Community detection via measuring the strength between nodes for dynamic networks
Gaumont et al. Finding remarkably dense sequences of contacts in link streams
Deylami et al. Link prediction in social networks using hierarchical community detection
Nedioui et al. Detecting communities in social networks based on cliques
Gao et al. Link prediction based on linear dynamical response
Ilhan et al. Community event prediction in dynamic social networks
Mahmood et al. Measuring scientific collaboration in co-authorship networks
Yun et al. Return to basics: Clustering of scientific literature using structural information

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant