CN104217073B - A kind of visual layout's method of network community gravitation guiding - Google Patents
A kind of visual layout's method of network community gravitation guiding Download PDFInfo
- Publication number
- CN104217073B CN104217073B CN201410425693.9A CN201410425693A CN104217073B CN 104217073 B CN104217073 B CN 104217073B CN 201410425693 A CN201410425693 A CN 201410425693A CN 104217073 B CN104217073 B CN 104217073B
- Authority
- CN
- China
- Prior art keywords
- corporations
- node
- center
- gravitation
- layout
- 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.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
It is specifically a kind of to reflect visual layout's method of complex network community structure the present invention relates to information visualization field.The present invention adds corporations' gravitation on the basis of power guiding placement algorithm to each node, the node of same corporations is gathered to the center of corporations, k means algorithms are introduced into corporations' gravitation, realizes the cluster that node is completed while layout.It is overlapping in order to prevent and reduce, introduce the size of gravitational coefficients control corporations of corporations gravitation.The present invention can not only show the community structure of complex network, also have it is simple, be easily achieved and the features such as fast convergence rate.
Description
Technical field
It is specifically a kind of to reflect the visualization cloth of complex network community structure the present invention relates to information visualization field
Office's method.
Background technology
Visual layout's method is widely used in social relation network field, can be for pass between displaying individual and individual
The tightness degree of system, displaying spreading network information process directly perceived etc..Current, in visual layout's method most commonly firmly
Guiding (force-directed) method shows network structure, and achieves good analytical effect.But, power guiding side
Method does not account for graph-clustering phenomenon mostly, and its node is uniformly distributed this layout principles, prevents network clustering feature
Show.
It can not mostly show that complex network has this characteristic of community structure to solve power guiding visual layout method,
Researchers are by introducing community detecting algorithm, it is proposed that the method for first network clustering visual layout again shows complex network
Community structure.Zhu Zhi it is good et al.《CAD and figure journal》Entitled " base has been delivered on the o. 11th of volume 23
The article of the network topology structure visual layout algorithm divided in complex network community ", this article proposes a kind of based on corporations
The network topology algorithm of division.Find that algorithm carries out corporations' division to the node in network first with complex network community, and
By a corporations it is abstract be a node, new network is built with the side that is associated as between corporations;On this basis, with physics class
Ratio method determines the position of corporations' central point, and determines according to the scale of corporations the regional extent of corporations;Finally selected with condition
Excellent mode fills corporations' internal node to complete the layout of network topology.This layout method can enter to the community structure of network
Row is shown, but corporations are divided and placement algorithm phase separation, and layout result can not be carried out effectively to fringe node
Displaying.Publication No. CN101741623A Chinese invention patent discloses a kind of network visualization side of network technique field
Method, employs the group dividing method based on modularity index and carries out layer division to network, also depends on existing corporations and draws
Point method is first clustered to network, is then laid out again.So that layout step is excessively complicated, there is positioning efficiency low scarce
Point.
Therefore, first network clustering again visual layout method show network community structure when there is following ask
Topic:(1) layout method too relies on the community detecting algorithm existed;(2) layout result subjective factor is too many, some nets
Network structural information has lost;(3) algorithm steps are complicated, are not suitable for catenet.
The content of the invention
The technical problems to be solved by the invention be in order to solve the limitation of first network clustering visual layout's method again,
The displaying of complex network community structure is completed while realizing layout, a kind of visual layout side of corporations' gravitation guiding is proposed
Method.
This method is on the basis of power guiding layout, to add corporations' gravitation to each node, and introduce k-means
Algorithm principle, enables the node of same corporations to gather to the center of corporations.Corporations are controlled to draw by corporations' gravitational coefficients
The size of power, prevents and reduces overlapping, has reached good layout effect.For corporations' gravity-based clustering, according to the important of node
Degree is that node sets quality, then asks for the centre distance of node corporations belonging to, finally sets rational corporations' gravitation system
Count to obtain corporations' gravitation of each node.Node is gathered under the guiding of corporations' gravitation to the center of corporations.Specific method
Including:
A kind of visual layout's method of corporations' gravitation guiding, is that node determines quality according to node center degree;According to society
The nodes of group, the position of node determines the center of corporations, and node chooses institute according to it to the distance of corporations center
Belong to corporations;The node of other in calculating network is to the gravitation and repulsion of node, according to gravitation and repulsion to node layout;Calculate corporations
To corporations' gravitation of node, node is gathered under the guiding of corporations' gravitation to the center of corporations so that the node of same corporations gathers
Gather together, realize the effect of cluster.
According to the actual range d and ideal distance r of two nodes, formula is called:Calculating comes from
The gravitation f on sidea(d) with the repulsion f for coming from noder(d);According to node v mass M [v], node v to corporations CkCenter
ukApart from dk, call formula:fg(d)=gM [v] min (d1、d2、...、dk) corporations gravitation of the corporations to node v is calculated, its
In, g is corporations' gravitational coefficients.
Node layout is influenceed by three active forces, is gravitation, all nodes between the node for having side connected respectively
Between repulsion and all nodes and affiliated corporations corporations' gravitation.For the determination of corporations center, k-means is introduced
Algorithm asks for corporations center.Determine that corporations center specifically includes step:(1) k node is randomly selected as initial corporations
Center;(2) the initial corporations center nearest with nodal distance is chosen as the corporations center of the node, and node-home is in the society
Group;(3) after corporations' ownership of all nodes is determined, formula is calledDetermine the center of k-th of corporation
Position, n represents corporations k nodes, pvRepresent node v position;(4) repeat step (2)-(3) determine the center of each corporations
Position, when system temperature reaches minimum value, all corporations centers are determined.
For preventing and reducing overlapping, it is necessary to be realized by the size for adjusting corporations' gravitational coefficients.Generally, node
Number is more, and corporations' gravitational coefficients are smaller.By experiment, the values of corporations' gravitational coefficients is optimal general to be taken between 0-2.
Network clustering is combined by the present invention with visual layout, improves the objectivity of layout result, it is therefore prevented that edge
The loss of nodal information, and algorithm is simple, be easily achieved with fast convergence rate etc..
Brief description of the drawings
Visual layout's method flow diagram of Fig. 1 corporations gravitation guiding;
Layout effect of the layout of Fig. 2 present invention on Zachary data sets.
Embodiment
Present disclosure is described in further detail below in conjunction with the accompanying drawings.
Whole system will be issued to balance in repulsion, gravitation and corporations' gravitational interaction, and wherein corporations' gravitation is used to draw
Node is led to corporations center to gather.Secondly, it is not directed through carrying out node corporations' division obtaining corporations center, but
K-means algorithm principles are introduced into corporations' gravitation, the determination at corporations center is completed.Node overlapping problem is finally directed to, to society
Group's gravitational coefficients are adjusted, and prevent node from gathering to corporations' central point transition.
The quality of node each first is determined by the importance of node.Node center degree that can be in network analysis
(degree centrality), close centers degree (closeness centrality) and spacing centrad (betweenness
Centrality) as the standard for weighing pitch point importance, the layout that the different quality criteria for classifying is produced can show net
Network Clustering features.Secondly, node is determined to the distance at corporations center, and the generation of Clustering Effect is mainly determined by the distance.Then
Corporations' gravitation size of each node is obtained according to corporations' gravitational coefficients.The value of corporations' gravitational coefficients depends primarily on node
Quantity, in general, nodes are more, and corporations' gravitational coefficients value is smaller.
If G is a network, G (V, E) is expressed as with node and side, wherein V is the set { v of n node1,v2,...,
vn, E is the set { e on m bars side1,e2,...,em, G can be divided into k corporations { C1,C2,…,Ck, in corresponding corporations
The heart is { u1,u2,…,uk}.If location problem to be equivalent to the stressing conditions of object in physics, then node will be in three power
In the presence of reach balance:Come from the repulsion of other nodes, the corporations of corporations draw where coming from the gravitation on side and coming from
Power.Specifically implementation steps are as shown in Figure 1:
A1:For initial phase.Node center degree method in network analysis determines quality for node, i.e., accordingly in
The quality of the bigger node of heart degree is bigger.
A2:Gravitation suffered by calculate node and repulsion, gravitation and repulsion are mainly used to the balance of maintenance system and reduce side friendship
Fork.According to the actual range d and ideal distance r of two nodes, the FR algorithms of formula (1) are called to calculate gravitation fa(d) with repulsion fr
(d).According to gravitation fa(d) with repulsion fr(d) to node layout, the principle of layout be there is the node on side should be adjacent, but will
Maintain a certain distance, optimum distance depends on the quantity of node and the size of painting canvas.Repulsion is to be present in all nodes, is drawn
Power exists only in the node on side.The moving range of layout iteration node can be with the reduction reduced progressively of temperature every time
System temperature is preferably minimized, and layout can also reach optimum state.
A3:Determine corporations center.Corporations center is asked for using based on k-means algorithms.Specific corporations centre bit
Put determination method as follows:(1) K node is randomly selected as initial corporations center, and K represents corporations' number;(2) other sections are asked for
Point and the distance at this K corporations center, choose the initial corporations center nearest with nodal distance as the corporations center of the node,
Node-home is in the corporations;(3) after corporations' ownership of all nodes is determined, the center of each corporations is determined, is called
FormulaThe center of k-th of corporation is determined, n represents corporations k nodes, pvRepresent node v position;(4)
Repeat step (2)-(3), when system temperature reaches minimum value, all corporations centers are determined.
A4:According to corporations gravitation fg(d) further adjustment is laid out so that the node of same corporations can flock together,
Realize the effect of cluster.According to node v mass M [v], node v to corporations central point ukApart from dk, call formula (2) to calculate
Come from corporations gravitation of the place corporations to node v, wherein, g is corporations' gravitational coefficients, dkFor node v and corporations k distance:
fg(d)=gM [v] min (d1、d2、...、dk) (2)
A5:Corporations gravitational coefficients g value is adjusted, so as to prevent node transition from gathering.G's depends mainly on the size of node
Quantity, in general, nodes are more, and g values are smaller.Its value is generally between 0 to 2.
A6:When the temperature of system reaches given minimum value, knot adjustment terminates, and otherwise performs step A2.For temperature
Adjustment, simulated annealing principle can be used, is the first high temperature value of a given comparison, then reduction this temperature slowly
Value, until being minimized value.
In order to quantify the degree of strength of the community structure produced under the condition of convergence, introduce modularity Q and assess community structure
Power.
Wherein, kiAnd kjRepresent node i and the j number of degrees, Ci and Cj represent node i and the corporations belonging to j, and m is network G
Total side number.When i and j belong to same corporations, δ (CiCj)=1, otherwise δ (CiCj)=0.AijRepresent i and j connection feelings
Condition, A when there is side between i and jij=1, otherwise Aij=0.Modularity has been widely used among corporations' detection, it
Value is between 0 to 1, typically when module angle value is more than 0.3, it can be found that community structure.
Layout effect of the layout of Fig. 2 present invention on Zachary data sets.Public data collection Zachary is have selected to enter
Row experiment.Zachary networks describe the mutual social relationships between the karate clubbite of university of one, the U.S., include 34
Individual node, 78 sides.Node on behalf clubbite, while mutually recognizing between representing two members.It is laid out effect such as Fig. 2 institutes
Show, it can be seen that the algorithm can substantially reflect the community structure characteristic of complex network.
Claims (3)
1. a kind of visual layout's method of corporations' gravitation guiding, it is characterised in that:It is that node determines matter according to node center degree
Amount;According to corporations' nodes, the position of node determines the center of corporations, and node arrives the distance of corporations center according to it
Corporations belonging to choosing;According to the node and the actual range d and ideal distance r of other nodes, formula is called:Calculating comes from gravitation f of the side to the nodea(d) and repulsion f of other nodes to the node is come fromr
(d);According to gravitation and repulsion to node layout;According to node v mass M [v], node v to corporations CkCenter ukDistance
dk, call formula:
fg(d)=gM [v] min (d1, d2..., dk) corporations gravitation of the corporations to node v is calculated, wherein, g is gravitation system of corporations
Number;Node is gathered under the guiding of corporations' gravitation to the center of corporations so that the node rendezvous of same corporations together, is realized
Cluster.
2. according to the method described in claim 1, it is characterised in that repulsion is present between all nodes, and gravitation is existed only in
Have between the node on side, the principle of layout is:The node that there is side is adjacent, and node is gathered to the center of affiliated corporations, system temperature
Degree is preferably minimized, and layout is completed.
3. method according to claim 1 or 2, it is characterised in that determine that corporations center specifically includes step:(1)
K node is randomly selected as initial corporations center;(2) choose the initial corporations center nearest with nodal distance and be used as the node
Corporations center, node-home is in the corporations;(3) after corporations' ownership of all nodes is determined, formula is calledThe center of k-th of corporation is determined, wherein, n represents corporations k nodes, pvRepresent node v position;
(4) repeat step (2)-(3) determine the center of each corporations, when system temperature reaches minimum value, all corporations centers
It is determined that.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410425693.9A CN104217073B (en) | 2014-08-26 | 2014-08-26 | A kind of visual layout's method of network community gravitation guiding |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410425693.9A CN104217073B (en) | 2014-08-26 | 2014-08-26 | A kind of visual layout's method of network community gravitation guiding |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104217073A CN104217073A (en) | 2014-12-17 |
CN104217073B true CN104217073B (en) | 2017-07-25 |
Family
ID=52098559
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410425693.9A Active CN104217073B (en) | 2014-08-26 | 2014-08-26 | A kind of visual layout's method of network community gravitation guiding |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104217073B (en) |
Families Citing this family (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105654387B (en) * | 2015-03-17 | 2020-03-17 | 重庆邮电大学 | Time-varying network community evolution visualization method introducing quantization index |
US10269152B2 (en) | 2015-06-05 | 2019-04-23 | International Business Machines Corporation | Force-directed graphs |
CN105260611B (en) * | 2015-10-26 | 2018-11-13 | 浙江工业大学 | Figure layout method based on nodal community transmission function |
CN106708839B (en) * | 2015-11-12 | 2021-04-23 | 创新先进技术有限公司 | Relational network architecture display method and device |
CN106919584B (en) * | 2015-12-26 | 2020-07-07 | 华为技术有限公司 | Topological graph layout method and device |
CN105912562B (en) * | 2016-03-21 | 2019-05-14 | 华南理工大学 | Data visualization method and its system based on hierarchical model |
CN105959132B (en) * | 2016-04-21 | 2019-09-10 | 重庆邮电大学 | A kind of community's layout method for visualizing towards large scale network |
CN106484804A (en) * | 2016-09-22 | 2017-03-08 | 北京金堤科技有限公司 | Key element based on electric field simulation and relation methods of exhibiting |
CN106682811B (en) * | 2016-11-23 | 2020-11-03 | 广西中烟工业有限责任公司 | Market network visualization method based on density clustering and force guiding algorithm |
CN106709507B (en) * | 2016-11-29 | 2019-11-08 | 北京林业大学 | A kind of parallel coordinate system view cluster data binding method of power guiding segmentation bone |
CN106934422B (en) * | 2017-03-16 | 2019-07-26 | 浙江工业大学 | Hierarchical visual abstraction method based on improved force guide diagram layout |
CN108319904A (en) * | 2018-01-24 | 2018-07-24 | 江苏金陵科技集团有限公司 | A kind of virtual interactive interface method assisting face or Car license recognition |
CN109245921A (en) * | 2018-08-24 | 2019-01-18 | 电子科技大学 | A kind of network topology structure method for visualizing in irregular area |
CN109766478B (en) * | 2019-01-08 | 2021-06-29 | 浙江财经大学 | Semantic-enhanced large-scale multivariate graph simplified visualization method |
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 |
CN111666312B (en) * | 2020-05-13 | 2023-05-23 | 中国科学院软件研究所 | Social network community cluster mining method and system based on universal gravitation law |
CN112507018A (en) * | 2020-12-08 | 2021-03-16 | 南京理工大学 | Information presentation method based on gravity model |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102893275A (en) * | 2010-05-14 | 2013-01-23 | 微软公司 | Automated social networking graph mining and visualization |
-
2014
- 2014-08-26 CN CN201410425693.9A patent/CN104217073B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102893275A (en) * | 2010-05-14 | 2013-01-23 | 微软公司 | Automated social networking graph mining and visualization |
Non-Patent Citations (1)
Title |
---|
《BBS虚拟社区的演化规律探索及仿真》;吴渝 等;《系统工程理论与实践》;20101031;第30卷(第10期);第1883-1890页 * |
Also Published As
Publication number | Publication date |
---|---|
CN104217073A (en) | 2014-12-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104217073B (en) | A kind of visual layout's method of network community gravitation guiding | |
CN103106279B (en) | Clustering method a kind of while based on nodal community and structural relationship similarity | |
CN109167835A (en) | A kind of physics resource scheduling method and system based on kubernetes | |
CN110366193B (en) | Slice deployment method and device for network arrangement layer bearing of smart grid | |
CN106446959A (en) | Method and device for dynamically matching cloud computing resources | |
CN112464784A (en) | Distributed training method based on hybrid parallel | |
CN105260414B (en) | User behavior similarity calculation method and device | |
CN106934497A (en) | Wisdom cell power consumption real-time predicting method and device based on deep learning | |
CN104320353B (en) | A kind of method and apparatus of station address management | |
CN104935476B (en) | A kind of network traffics matrix measuring method based on SDN | |
CN105959132B (en) | A kind of community's layout method for visualizing towards large scale network | |
CN102122434A (en) | Urban public transport network optimization method capable of improving overall transfer performance | |
CN107659973A (en) | Super-intensive network cluster dividing method based on density K means algorithms | |
CN109508748A (en) | A kind of clustering method and device | |
CN106817314A (en) | Big data acquisition method, device and system | |
CN107862404A (en) | Manufacturing service supply chain optimization method based on service relevance | |
CN109117437A (en) | A kind of image feature extraction method towards image of clothing retrieval | |
CN103595652B (en) | The stage division of QoS efficiency in a kind of powerline network | |
CN105955710B (en) | Parallel artificial data processing method based on tree-like communication structure | |
CN107196808A (en) | A kind of method of double-layer network model construction | |
CN106789286A (en) | The display methods and device of a kind of network topological diagram | |
CN108171538A (en) | User data processing method and system | |
CN109685279A (en) | A kind of Complicated Distribution Network PQM optimization method based on topology degradation | |
CN103516792A (en) | Data deploying method based on network delay, network transmission energy consumption and server energy consumption | |
CN107944553A (en) | A kind of method for trimming and device of CNN models |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |