CN108287866A - Community discovery method based on node density in a kind of large scale network - Google Patents
Community discovery method based on node density in a kind of large scale network Download PDFInfo
- Publication number
- CN108287866A CN108287866A CN201711371510.XA CN201711371510A CN108287866A CN 108287866 A CN108287866 A CN 108287866A CN 201711371510 A CN201711371510 A CN 201711371510A CN 108287866 A CN108287866 A CN 108287866A
- Authority
- CN
- China
- Prior art keywords
- node
- community
- density
- com
- network
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/90335—Query processing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to the community discovery method based on node density in a kind of large scale network, including sets the topological structure of input network and indicated with figure G={ V, E }, and V and E are respectively the set on node and side;The averag density of calculate node density and whole network;All node density data are arranged and obtained density sequence, the density for spending maximum node is determined as network threshold;Community is found from the node in sequence, forms the community's sequence determined in network, and remaining node is constituted and is gathered;Judge set whether be remaining community set and isolated point set;Determine community's quantity of whole network;Optimization is merged to community.Node density method using the present invention, which carries out community discovery, lower computation complexity, and the accuracy that community divides is high, is suitable for the complex network of various scales and type.
Description
Technical field
The present invention relates to network community discovery methods, and in particular to the community based on node density in a kind of large scale network
It was found that method.
Background technology
At present community structure find have become complex network research in a hot issue, in recent years by computer,
The extensive concern of the area researches person such as mathematics, biology and sociology, such as the mapping knowledge domains of ongoing research area just combine
The theory of community discovery.And the presence of community structure is found that in these networks, the discovery of complex network community structure
There is important theory significance and practical value for the Analysis of Topological Structure of complex network, functional analysis and behavior prediction.It is multiple
Community in miscellaneous network is one group of set that is similar to each other and being constituted with other nodes have differences in network node, same society
Area's internal node interconnection is intensive, and intermediate node interconnection in community's is relatively sparse.
Community is the subset of nodes, and the connection between the interior nodes of same community is closer, and between different communities
Node connection it is relatively sparse.Node in same community has similar function in a network, therefore this community is in net
There are one specific effects in network.The community structure in complex network is studied for analyzing the topological structure of complex network, understanding
The behavior that function possessed by network and prediction network may have has very important significance, and additionally has extensive
Application prospect.
In recent years, many community detection methods are proposed out in succession, these algorithms are substantially divided into three classes:It is based on
The Hierarchical Segmentation method of figure, the method based on cluster and the method based on optimization, wherein the method based on optimization obtained it is more next
More concerns.But current community discovery algorithm generally has division community's accuracy not high, needs priori, calculates complicated
Spend the shortcomings that height leads to not apply in actual complex network.
Invention content
It is not high the technical problem to be solved by the present invention is to divide community's accuracy in community discovery algorithm, need priori to know
Know, computation complexity excessively high the shortcomings that leading to not apply in actual complex network.
The technical solution that the present invention solves above-mentioned technical problem is as follows:Society based on node density in a kind of large scale network
Area finds method, includes the following steps:
S1, the topological structure for setting input network indicate with figure G={ V, E }, and V and E are respectively the set on node and side;
S2, calculate node density Density (vi k) and whole network averag density MeanD (G);
S3, descending arrangement is carried out to the node density data of all nodes and obtains density sequence Seq (G), degree maximum
The density of node be determined as network threshold Thr (G);
S4, the community Com (i) in network is found by sequence Seq (G), form community's sequence that network determines, and will remain
Remaining node constitutes set Remain (G);
S5, judge Remain (G) whether the set for being remaining community set { Com (i) } and isolated point, if so, into
Step S6, otherwise return to step S2;
S6, the community's quantity for determining whole network;
S7, optimization is merged to community.
The beneficial effects of the invention are as follows:The concept for proposing node density of the present invention, to measure the close of nodes
Degree, and with the big seed node of node density value proceed by breadth first search extend community, and with its surrounding neighbours density
Maximum node is combined as first unit community, and obtains final community by successive ignition and optimization and divide, society in network
The fringe node in area more easily determines which community belonged to, and the crossover node between community is most difficult to determination and belongs to which community, and
It is not that traditional community finds to proceed by breadth first search from Network Central Node, node density computational methods of the present invention are just
It is high to meet fringe node density, crossover node low density feature in community's in network, method of the invention can be without priori
The case where knowledge, can directly determine that node belongs to community's degree, meanwhile, node density method using the present invention carries out community
It was found that there is lower computation complexity, the accuracy that community divides is high, is suitable for the complex network of various scales and type.
Description of the drawings
Fig. 1 is general flow chart of the present invention;
Fig. 2 is the calculating schematic diagram of interior joint density of the present invention;
Fig. 3 is the calculating schematic diagram of R values in the present invention;
Fig. 4 is node density schematic diagram of the present invention in Zachary ' s karate club.
Specific implementation mode
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the present invention.
As shown in Figure 1, the community discovery method based on node density in a kind of large scale network, includes the following steps:
S1, the topological structure for inputting network is set as G={ V, E }, whole network is indicated with figure G={ V, E }, V and E difference
For the set of node and side.
S2, calculate node densityWith the averag density MeanD (G) of whole network.
Step S2 specifically includes following steps:
As shown in Fig. 2, in an embodiment of the present invention, Fig. 2 (a) is figure G={ V, E }, Fig. 2 (b) is subgraph G'.
S21, will be to scheme the node v of G={ V, E }iSubgraph G' is formed for the starting point point that hop count is k forward, calculate node is close
Degree
In formula (1), i is node serial number, and k is with vi(had according to heterogeneous networks different for the forward direction hop count of starting point
Value), V' is the set of subgraph G' interior joints, | V'| is the quantity of V' interior joints, and E' is the set on side in subgraph G', | E'| is
The quantity on side in E'.
S22, pass through node densityCalculate the averag density MeanD (G) of whole network:
In formula (2), N is the quantity of G interior joints.
S3, density sequence Seq (G) is arranged the density data of node and is obtained, the density for spending maximum node
It is determined as network threshold Thr (G).
Step S3 specifically includes following steps:
S31, all nodes in network are arranged by the sequence of its density descending, when the density of node is identical, is pressed
Ascending order arrangement is carried out according to node serial number size, constructs the node density sequence Seq (G) of whole network.
The maximum node of S32, degree of finding out, and using its node density as network threshold Thr (G), when presence is maximum
When the case where node more than one, the node of density minimum in these nodes is found out, using the density of the node as network threshold
Thr(G)。
S4, community Com (i) is found from the node in sequence Seq (G), form community's sequence that network determines, and will remain
Remaining node constitutes set Remain (G).
Step S4 specifically includes following steps:
S41, first element in density sequence Seq (G) is chosen, the node representated by the element is as community Com
(1), breadth first search and since the node is carried out to network, the density value of the node searched meets Density (vi)<
Stop the search of its neighbor node when Thr (G), obtains neighbor node set N (1)={ ni}。
S42, traversal N (1), when the density value of set N (1) interior joint meets niWhen >=Thr (G), ni∈Com(1)。
S43, step S41 and step S42 is repeated, the node that the element in density sequence Seq (G) is represented constitutes community's sequence
Row:Com (1), Com (2) ..., Com (k), remaining node constitute residue node set Remain (G), and Remain (G) includes surplus
Remaining community's set { Com (i) } and isolated node rvj, j is the number of node.
S5, judge Remain (G) whether the set for being remaining community set { Com (i) } and isolated point, if so, into
Otherwise step S6 returns to step S2.
S6, the community's quantity for determining whole network.
Step S6 is specifically included:
Each community in { Com (1), Com (2) ..., Com (k) } ∩ { Com (i) } sequence is calculated according to formula (2) and is put down
Equal density meanD (i), as meanD (i) >=MeanD (G), which is community content COM (i), and otherwise, which is it
He is community com (j), is merged into optimised in community content COM (i), the community content COM (i) and other communities com
(j) quantity and be community's quantity in whole network.
S7, optimization is merged to community.
Step S7 specifically includes following steps:
S71, the community com (j) combined to needs, the maximum community COM (K) of R values after selection merges with community Com (i),
The number of the fixed communities K, and community com (j) is added in COM (K), the iteration process, until having handled all
Until com (j).The computational methods of R values are:
In formula (3), as shown in figure 3, in the embodiment of the present invention, BinFor connect community's C internal nodes side quantity,
BoutFor the quantity on the side of connection community's C internal nodes and external node.
If S72, isolated node rvjOnly it is connected with a community, then the isolated node is incorporated to community, if isolated node rvj
It is connected with multiple communities, then is incorporated to multiple communities using the isolated node as overlay node.
The node of the present invention as shown in Figure 4 in Zachary ' s karate club (Zha Kali karates club)
Density, in the embodiment of the present invention, it is high to meet fringe node density, and it is low to intersect dot density for community in network, can facilitate determining section
Point community's degree.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention.
Claims (6)
1. the community discovery method based on node density in a kind of large scale network, which is characterized in that include the following steps:
S1, the topological structure for setting input network indicate with figure G={ V, E }, and V and E are respectively the set on node and side;
S2, calculate node densityWith the averag density MeanD (G) of whole network;
S3, descending arrangement is carried out to the node density data of all nodes and obtains density sequence Seq (G), the maximum section of degree
The density of point is determined as network threshold Thr (G);
S4, the community Com (i) in network is found by sequence Seq (G), form community's sequence that network determines, and residue is saved
Point constitutes set Remain (G);
S5, judge Remain (G) whether the set for being remaining community set { Com (i) } and isolated point, if so, entering step
S6, otherwise return to step S2;
S6, the community's quantity for determining whole network;
S7, optimization is merged to community.
2. the community discovery method based on node density in large scale network according to claim 1, which is characterized in that institute
It states step S2 and specifically includes following steps:
S21, the node v in G will be schemediAnd with viFor starting point forward k-hop point form subgraph G', calculate node density
In formula (1), i is node serial number, and k is with viFor the forward direction hop count of starting point, V' is the set of subgraph G' interior joints, | V'
| it is the quantity of V' interior joints, E' is the set on side in subgraph G', | E'| is the quantity on side in E';
S22, pass through node densityCalculate the averag density MeanD (G) of whole network:
In formula (2), N is the quantity of G interior joints.
3. the community discovery method based on node density in large scale network according to claim 1, which is characterized in that institute
It states step S3 and specifically includes following steps:
S31, the node density progress descending arrangement of all nodes in network is pressed when there is a situation where that node density is identical
Ascending order arrangement is carried out according to node serial number size, constructs the node density sequence Seq (G) of whole network;
The maximum node of S32, degree of finding out, and using its node density as network threshold Thr (G), when the maximum node of presence
When the case where more than one, the node of density minimum in these nodes is found out, using the density of the node as network threshold Thr
(G)。
4. the community discovery method based on node density in large scale network according to claim 1, which is characterized in that institute
It states step S4 and specifically includes following steps:
S41, first element in density sequence Seq (G) is chosen, using the node representated by the element as community Com (1),
And breadth first search is carried out to network since the node, the density value for searching node meets Density (vi)<Thr(G)
When stop its neighbor node search, obtain neighbor node set N (1)={ ni};
S42, traversal N (1), when the density value of set N (1) interior joint meets niWhen >=Thr (G), ni∈Com(1);
S43, step S41 and step S42 is repeated, the node that the element in density sequence Seq (G) is represented constitutes community's sequence:
Remaining point is constituted residue node set Remain (G) by Com (1), Com (2) ..., Com (k), and Remain (G) includes surplus
Remaining community's set { Com (i) } and isolated node rvj, j is the number of node.
5. the community discovery method based on node density in large scale network according to claim 2, which is characterized in that institute
Step S6 is stated to specifically include:To each community in { Com (1), Com (2) ..., Com (k) } ∩ { Com (i) } sequence according to formula
(2) averag density meanD (i) is calculated, as meanD (i) >=MeanD (G), which is community content COM (i), otherwise, should
Community be other community com (j), be merged into optimised in community content COM (i), the community content COM (i) and other
The quantity of community com (j) and be community's quantity in whole network.
6. the community discovery method based on node density in large scale network according to claim 1, which is characterized in that institute
It states step S7 and specifically includes following steps:
S71, the community com (j) combined to needs, the maximum community COM (K) of R values after selection merges with community Com (i), K is
The number of determining community, and community com (j) is added in COM (K), the iteration process, until having handled all com
(j) until.The computational methods of R values are:
In formula (3), BinFor the quantity on the side of connection community's internal node, BoutFor connection community's internal node and outside segments
The quantity on the side of point;
If S72, isolated node rvjOnly it is connected with a community, then the isolated node is incorporated to community, if isolated node rvjWith it is more
A community is connected, then is incorporated to multiple communities using the isolated node as overlay node.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711371510.XA CN108287866A (en) | 2017-12-18 | 2017-12-18 | Community discovery method based on node density in a kind of large scale network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711371510.XA CN108287866A (en) | 2017-12-18 | 2017-12-18 | Community discovery method based on node density in a kind of large scale network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108287866A true CN108287866A (en) | 2018-07-17 |
Family
ID=62832126
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711371510.XA Pending CN108287866A (en) | 2017-12-18 | 2017-12-18 | Community discovery method based on node density in a kind of large scale network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108287866A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109359199A (en) * | 2018-08-27 | 2019-02-19 | 平安科技(深圳)有限公司 | Fund manager's group dividing method, system, computer equipment and storage medium |
CN111274457A (en) * | 2020-02-03 | 2020-06-12 | 中国人民解放军国防科技大学 | Network graph partitioning method and storage medium |
CN112994933A (en) * | 2021-02-07 | 2021-06-18 | 河北师范大学 | Generalized community discovery method for complex network |
-
2017
- 2017-12-18 CN CN201711371510.XA patent/CN108287866A/en active Pending
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109359199A (en) * | 2018-08-27 | 2019-02-19 | 平安科技(深圳)有限公司 | Fund manager's group dividing method, system, computer equipment and storage medium |
CN111274457A (en) * | 2020-02-03 | 2020-06-12 | 中国人民解放军国防科技大学 | Network graph partitioning method and storage medium |
CN111274457B (en) * | 2020-02-03 | 2023-12-19 | 中国人民解放军国防科技大学 | Network graph segmentation method and storage medium |
CN112994933A (en) * | 2021-02-07 | 2021-06-18 | 河北师范大学 | Generalized community discovery method for complex network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108287866A (en) | Community discovery method based on node density in a kind of large scale network | |
CN106709037B (en) | A kind of film recommended method based on Heterogeneous Information network | |
CN106886524A (en) | A kind of community network community division method based on random walk | |
CN105183796A (en) | Distributed link prediction method based on clustering | |
CN106951524A (en) | Overlapping community discovery method based on node influence power | |
CN106599090A (en) | Structure centrality-based network community discovery method | |
CN103020163A (en) | Node-similarity-based network community division method in network | |
CN105574541A (en) | Compactness sorting based network community discovery method | |
CN111259206B (en) | Multiple influence matrix weighting method in directional weighting network | |
CN105868791A (en) | Multi-resolution community discovering method based on fuzzy clustering | |
Negara et al. | A review on overlapping and non-overlapping community detection algorithms for social network analytics | |
CN109614520B (en) | Parallel acceleration method for multi-pattern graph matching | |
CN107945037A (en) | A kind of social networks based on node structure feature goes de-identification method | |
CN109345403A (en) | A kind of organization overlapping community discovery method based on localized mode lumpiness | |
CN102819611B (en) | Local community digging method of complicated network | |
CN110445654A (en) | A kind of social networks multi-source rumour source tracing method and system based on community's division | |
CN108133426B (en) | Social network link recommendation method | |
CN112989526B (en) | Aviation network key node identification method based on kernel extreme learning machine | |
CN106127595A (en) | A kind of community structure detection method based on positive and negative side information | |
Kong et al. | Dynamic weighted heuristic trust path search algorithm | |
CN108052743B (en) | Method and system for determining step approach centrality | |
CN108319727A (en) | A method of any two points shortest path in social networks is found based on community structure | |
Chang et al. | Community detection with attributed random walk via seed replacement | |
CN109561498B (en) | Sensor node distributed positioning method based on improved Newton method | |
Friedrich et al. | Evolutionary algorithms and submodular functions: Benefits of heavy-tailed mutations |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180717 |