CN104933621A - Big data analysis system and method for guarantee ring - Google Patents
Big data analysis system and method for guarantee ring Download PDFInfo
- Publication number
- CN104933621A CN104933621A CN201510342957.9A CN201510342957A CN104933621A CN 104933621 A CN104933621 A CN 104933621A CN 201510342957 A CN201510342957 A CN 201510342957A CN 104933621 A CN104933621 A CN 104933621A
- Authority
- CN
- China
- Prior art keywords
- guarantee
- node
- relation
- function
- index
- 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
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention provides a big data analysis system for a guarantee ring. The big data analysis system comprises a work scheduling module, an analysis module and a service presentation module, wherein the work scheduling module is used for receiving a command of a user and coordinating the work of the analysis module and the service presentation module; the analysis module is used for analyzing guarantee data and at least acquiring an index which is relevant to the centrality of a guarantee node; and the service presentation module is used for presenting an analysis result of the analysis module in a certain form. The invention also provides an analysis method for finance guarantee risks.
Description
Technical field
The present invention relates to a kind of Risk Management for Financial Institutions system, particularly relate to a kind of large data analysis system and the method that can analyze and represent financial institution's guarantee business complex relationship.
Background technology
China's Guarantee originates from the end of the nineties in last century; first guarantee agency is distributed in developed area after occurring, day by day growing up but be badly in need of the medium-sized and small enterprises of credit aid is that Guarantee provides great market; particularly in the area of non-governmental capital richness, the demand of Guarantee is very strong.Along with the fast development of China's economy, guarantee industry also forms scale gradually, but over the past two years, because credit threshold improves in bank, and the clean-up and rectification of supervision department, shuffling greatly appears in guarantee industry, for guarantor enterprise, particularly private guarantor enterprise wants to develop and set up obstruction.Lack of standardization owing to managing, the high risk characteristic of industry in addition, one deck shade is shrouded in the development of guarantee industry all the time.Under the background of economic prosperity, the functioning condition of enterprise is generally good, and money collecting is guaranteed, assures that the risk of this pattern is relatively little.But when economic situation glides, the fund chain of guarantee corporation is just easy to because bad credit rises and tightens.
Make that market is helpless, the persistent ailment of order supervision layer headache with the guarantee circle credit risk of the complicated guarantee forms such as chain of rings guarantee, UNPROFOR, familial guarantee always, weakness one ring in the credit management of Ye Shi financial institution, has Risk-warning work to improving guarantee business management.How effectively announcement and the enterprise (guarantor and guarantee) involved by complex management guarantee circle are institute of current financial institutional risk administrative authority urgent problems.
Summary of the invention
In order to overcome above defect, the invention provides a kind of analytic system analyzed for financial guaranty insurance, comprising: scheduling module, analysis module and service display module; Described scheduling module is for the work of the instruction coordinating analysis module and service display module that receive user; Described analysis module is used for guarantee data analysis, and at least draws the index relevant to assuring the centrad of node; Described service display module is used for the analysis result of analysis module to show with certain forms.
Preferably, it is one or more that described service display module comprises in following functions unit: query unit, for inquiring about and specifying the guarantee information assured node or specify colony relevant; Form unit, assures node for providing by the needs of user or specifies the form that colony is relevant to specifying; Visualization, for realizing whole guarantee network and assuring the visual of colony's localized network.
Preferably, also comprise: data memory module, for store the raw data of input system, analysis module and service display CMOS macro cell process after data.
It is preferably, described that to comprise to the relevant index of centrad of guarantee node in the out-degree of assuring node and in-degree index, centrality degree index of social network of overall importance, locality centrality degree index of social network one or more.
Preferably, described out-degree and in-degree index are calculated by Degree function.
Preferably, described centrality degree index of social network of overall importance is calculated by Eigenvector Centrality function.
Preferably, described locality centrality degree index of social network is calculated by Pagerank function.
Preferably, described analysis module carries out level calculating by enumerative technique to guarantee hoop net network, enumerates from a node to all guarantee path of another one node in guarantee network, and enumerates guarantee incidence relations all between any two nodes.
Preferably, when described analysis module carries out level calculating, the basis for selecting of start node is centrality degree index of social network of overall importance or the locality centrality degree index of social network of node, or random selecting.
Preferably, described analysis module carries out maximum level scope when level calculates is 3 ~ 8.
Preferably, guarantee relation is categorized into several basic form by the result that described analysis module calculates according to level.
Preferably, described several basic form comprises man-to-man unidirectional guarantee relation, man-to-man mutual guarantor's relation, three above colonies of client protect in relation, guarantee chain relation mutually one or more.
Preferably, the result that level calculates by described analysis module is processed by Modularity function, thus guarantee relation is categorized into several basic form.
Preferably, described Modularity function uses Modularity Maximization method.
The present invention also provides a kind of analytical approach of financial guaranty insurance, comprising: scheduling step, receives the instruction of user and coordinates the work of service display module and analysis module; Analytical procedure, to guarantee data analysis, and at least draws the index relevant to assuring the centrad of node; Result shows step, is shown by analysis result with certain forms.
Preferably, described result show step comprise the following steps in one or more: query steps, inquire about and feed back and guarantee information that appointment is assured node or specified colony relevant; Report generation step, provides by the needs of user and assures node or specify the form that colony is relevant to specifying; Visualization processing step, carries out visual presentation to whole guarantee network and guarantee colony localized network.
Preferably, also comprise: data storing steps, for the data after stores processor.
It is preferably, described that to comprise to the relevant index of centrad of guarantee node in the out-degree of assuring node and in-degree index, centrality degree index of social network of overall importance, locality centrality degree index of social network one or more.
Preferably, described out-degree and in-degree index are calculated by Degree function.
Preferably, described centrality degree index of social network of overall importance is calculated by Eigenvector Centrality function.
Preferably, described locality centrality degree index of social network is calculated by Pagerank function.
Preferably, by enumerative technique, level calculating is carried out to guarantee hoop net network, enumerate from a node to all guarantee path of another one node in guarantee network, and enumerate guarantee incidence relations all between any two nodes.
Preferably, the basis for selecting carrying out the start node of level calculating is centrality degree index of social network of overall importance or the locality centrality degree index of social network of node, or random selecting.
Preferably, the maximum level scope of carrying out level calculating is 3 ~ 8.
Preferably, guarantee relation is categorized into several basic form by the result calculated according to level.
Preferably, described several basic form comprises man-to-man unidirectional guarantee relation, man-to-man mutual guarantor's relation, three above colonies of client protect in relation, guarantee chain relation mutually one or more.
Preferably, the result that level calculates is processed by Modularity function, thus guarantee relation is categorized into several basic form.
Preferably, described Modularity function uses Modularity Maximization method.
The guarantee large data analysis system of circle provided by the invention and method, based on hierarchical algorithms method, various centrad algorithm and grouping method synergistic application are got up, achieve the analysis of group feature, stratification, chain-type, be analyze existing guarantee circle only to break through based on the ability of guarantee client and direct securities relationship analysis, clearly can build all guarantor enterprises in financial institution, by the multi-layer guarantee relation between guarantor enterprise; Four kinds of grown forms of guarantee network are proposed, and realize to this four classes grown form carry out differentiated statistic of classification, represent, the guarantee form of various complexity is effectively identified and classified, so that help financial institution according to different risk class guarantee circle involved by guarantee and carried out Group management by guarantor enterprise client, really accomplish, because plan executed by circle, to solve the problem of generalization of traditional analysis, localization, one-sided; By calculating the centrad etc. of client, can determine to belong to the guarantee of network center of gravity/by guarantor enterprise in guarantee hoop net network, thus helping financial institution's quick position to need the Very Important Person carrying out Risk-warning and management and control; Propose the Very Important Person come with degree, pagerank, Eigenvector Centrality in fixer network, solve tradition guarantee and analyze the problem lacking and guarantee client is lacked to evaluation index.
Accompanying drawing explanation
Large data analysis system structural drawing is enclosed in the guarantee of Fig. 1 involved by embodiment of the present invention;
Large data analysis system processing flow chart is enclosed in the guarantee of Fig. 2 involved by embodiment of the present invention;
The Ntree function of Fig. 3 involved by embodiment of the present invention enumerates the processing flow chart of the relation between client.
Embodiment
Illustrated embodiment sets forth the present invention with reference to the accompanying drawings below.This time disclosed embodiment can be thought and is illustration in all respects, and tool is not restricted.Scope of the present invention not limit by the explanation of following embodiment, only by shown in the scope of claims, and comprises and to have same looking like and all distortion in right with right.
For solving the problems of the technologies described above, client's importance stratification in guarantee relation, by various centrad algorithm and grouping method synergistic application, shows by the present invention, thus comprehensively, Systematic Analysis guarantee relation, take precautions against guaranty insurance.
Concrete, the invention provides a kind of large data analysis system assuring circle, this system comprises business diagnosis module, Data import and memory module, function computation and analysis module and scheduling module.Data import and memory module are used for storing from the next guarantee relation data of banking system collection and some customer profile data; Function computation and analysis module is formed primarily of analytic function, includes client's importance valuation functions, hierarchical structure function, colony's partition function.Business diagnosis module is report capability and graphical representation function mainly.Scheduling module can dispatch the work of other modules.Wherein, function computation and analysis module is the core of native system, comprise statistical analysis unit and three class applied analysis functions to realize guarantee circle and analyze: the first kind is assessment guarantee node importance in network and central function, comprises Degree function (calculating the degree centrality of guarantee network node), Eigenvector Centrality function (calculating the eigenvector centrality degree of guarantee network node), Pagerank function (calculating the pagerank of guarantee network node); Equations of The Second Kind is the function in the guarantee path of summing up between whole guarantee network client, utilizes Ntree function to realize; 3rd class is the function being divided into various guarantee colony by statistical study and clustering algorithm, can accomplish that group feature manages to guarantee user.
Below in conjunction with embodiment, the large data analysis system of guarantee of the present invention circle is described.
Fig. 1 is that large data analysis system structural drawing is enclosed in the guarantee that embodiment of the present invention relates to.This system comprises: business diagnosis module 1, Data import and memory module 2, function computation and analysis module 3 and scheduling module 4.
Wherein, business diagnosis module 1 comprises query unit 11, form unit 12 and visualization 13.Wherein, query unit 11 is directly connected with memory module 2 with Data import, as a given client, can inquire about relative institute secured relation and guarantee index; When a given colony, institute's secured relation and the amount of guarantee etc. related in colony can be inquired about.Form unit 12 is directly connected with memory module 2 with Data import, not only can provide client's level form of the guarantee number of users, amount of guarantee etc. that each client is associated with for not at the same level time; Also colony's level form of the guarantee character of each colony, client's number, amount of guarantee etc. can be provided; The time form of the change (comprising the change of number of users, colony's number, the amount of money etc.) of different times guarantee relational network can also be provided.Visualization 13 is directly connected with memory module 2 with Data import, and what realize whole guarantee network and guarantee colony localized network is visual, can represent guarantee colony guarantee and by guarantee relation.
Data import and memory module 2 store the data model of the required data loaded, namely this Module Specification has the form (if the add-in in data model form is data layout or type, data content, load field, loading record number, Data Source etc.) of the requirement loading or store data, makes other unit of system or module can transfer required data information more easily; This module can load and the application data needed for storage system, and this module can store parameter and the data of each unit calculating simultaneously, and this module can also provide the SQL of data storing platform to calculate.
Function computation and analysis module 3 comprises statistical analysis unit 31 and power function unit 32.Wherein, statistical analysis unit 31 can carry out statistical induction to the result that each function of power function unit 32 calculates, by guarantee relation in a pair one unidirectional guarantee relation, man-to-man mutual guarantor's relation, three above colonies of client protect relation, basic form that guarantee chain relation four kinds is different mutually; Statistical study can also be carried out to the projects such as number, amount of guarantee, contract number that comprise of this several grown form.Power function unit 32 comprises Degree function 321, Eigenvector Centrality function 322, Pagerank function 323, Ntree function 324, Modularity function 325.Wherein, Degree function 321 be used for calculating guarantee network each node guarantee client number and by guarantee client number; Degree is made up of out-degree Outdegree and in-degree Indegree two centrality degree index of social network, and out-degree Outdegree refers to that the client in guarantee circle assures other people number; In-degree Indegree refers to the number that in guarantee circle, client tenders guarantee for it, as long as Outdegree and Indegree can obtain through simple demographics.Eigenvector Centrality function 322 is used for calculating guarantee network each node diagnostic vector center degree, this is the Graph Analysis algorithm of standard, can be used for weighing guarantee node users center degree in the entire network, its centrad evaluation is the evaluation of overall importance of whole network, the Network Science introduction that algorithm can be write with reference to Wang little Fan, Li Xiang, Chen Guanrong " (Beijing: Higher Education Publishing House, 2012: P165).Pagerank function 323 is used for calculating the pagerank assuring each node of network, be used for weighing guarantee node users importance in the entire network, its Assessment of Important is locality, the i.e. main impact by adjacent node, pagerank is google dedicated algorithms, invented by Larry Page and Sergey Brin, algorithm can refer to its paper " The PageRank Citation Ranking:Bringing Order to the Web ".Ntree function 324 carries out level calculating to guarantee hoop net network, is used for enumerating in guarantee network, from a node to all possible guarantee path of another one node, can arrange the maximum level time in path in function; Based on the level in this guarantee path, guarantee incidence relations all between any two clients can be delineated.Guarantee network can be carried out being divided into the colony independently with guarantee incidence relation by Modularity function 325, the Modularity Maximization algorithm that these function computing method adopt M.E.J. Newman to deliver on PNAS in 2006.Above Degree function 321, Eigenvector Centrality function 322, Pagerank function 323, Ntree function 324, Modularity function 325 trigger by statistical analysis unit 31, these functions read data from Data import and memory module 2, carry out computational analysis, the result after calculating returns and is stored in Data import and memory module 2.
Scheduling module 4 comprises input, the output of system, the output of the result such as form in the input of instruction and business diagnosis module can be carried out, dispatch other each modules of system or unit by this module operation personnel are manual, system can also carry out scheduling to other modules of system or unit automatically.
The analytic functions such as Degree function 321, Eigenvector Centrality function 322, Pagerank function 323, Ntree function 324, Modularity function 325 adopt Java analytic function bag and MapReducec mode to develop, and these functions can be called by SQL statement.Statistical analysis unit 31 is realized by SQL script.Query unit 11, form unit 12 two are developed report program with JSP and are realized Flexible Query and report capability.Scheduling module 4 adopts Java to develop.Whole system adopts B/S framework, and Data import and memory module 2 adopt SQL, and application program adopts Java, and network and visualization 13 adopt JSP programming language to realize exploitation.
Large data analysis system processing flow chart is enclosed in the guarantee of Fig. 2 involved by embodiment of the present invention.
First, start the large data analysis system of guarantee circle, according to operational needs, determine to need which data, now whole system enters duty, operating personnel can carry out the scheduling (step S1) of work by scheduling module 4.Loaded by the loading of scheduling module 4 data dispatching and memory module 2 and the data of storage service needs, the data of loading comprise all information (step S2) such as the data of guarantee network Water demand, and operational contract, promise breaking, refund.Concrete, there is the data model of the data required for analyzing in Data import and memory module 2, as long as put into data to corresponding table according to the requirement of data model, corresponding analysis can be started.
After Data import and memory module 2 load corresponding data, just the scoring of centrad can be carried out to each client, scoring process is started by scheduling module 4, statistical analysis unit 31 triggers Degree function 321, Eigenvector Centrality function 322, Pagerank function 323, calculate the guarantee network data of whole loading respectively, generate the centrad evaluation index of each client in guarantee circle, be respectively indegree, outdegree, eigenvectorcentrality and pagerank tetra-evaluation indexes, wherein Eigenvector Centrality and PageRank needs to arrange 2 model parameters: maximum iteration time
kand least error
θ, determine that the principle of parameter needs the hardware configuration according to system deployment, network size, computational accuracy to require to determine, preferably
kbetween 10 ~ 100,
θbetween 0.1 ~ 0.001, more preferably,
k=30,
θ=0.01, can revise depending on data cases, the index result simultaneously calculated all is stored in Data import and memory module 2 (step S3).After obtaining these indexs, according to the needs of business, utilize the screening function that statistical analysis unit 31 provides, all or part of according to above four indexs, the concern standard of system automatically setting center customer, obtains the client's list (step S4) needing to pay close attention to.
After having calculated the Score index in step S3, operating personnel can manual triggers Ntree function, and input data source can the level parameter K of Ntree, carry out level calculating by Ntree function and obtain the guarantee transmission path of any two clients, postrun result is exactly the guarantee chain that length is less than K, these chains enumerate all possible guarantee composition of relations, same guarantee client between may be present in many guarantee paths in, these data splittings are bases (step S5) of group clustering.Concrete, Ntree function enumerates the step of the relation between whole client (suppose there is N number of client) as shown in Figure 3.First, all clients of Water demand are numbered (suppose total N number of client) (S51), coding rule is as follows: customer number can be S according to centrad serial number from big to small according to Call center's degree evaluation indexes such as pagerank or the eigenvector centrality obtained in step S3
0, S
1... S
n-1, can be also S by whole client's random number
0, S
1... S
n-1.If the number parameter of client
i=0(step S52).Selected S
i for initial client (step S53).Then given maximum computation levels K, and set user's progression
j=0(step S54).Following calculating S
ij one-level ntree child node A
1, A
2... A
m; And judge S
ij with A
1..., A
mbetween whether form circulation guarantee, and record path circulation mark (step S55).Then by client
ilevel time
jadd 1:
j=
j+ 1(step S56).Judge client
ilevel time
jwith the size (step S57) of the computation levels K of setting, when
jit is yes for being less than K(step S57), then return step S55; If
jit is no for being more than or equal to K(step S57), then former initial client S
i middle customer number
ifrom adding 1, namely
i=
i+ 1(step S58).Judge the numbering of initial client and the size (step S59) of client's sum N, if
iit is yes for being less than N(step S59), then return step S53; If
iit is yes for being greater than N(step S59), then Ntree enumeration process terminates (step S510).The computing power of network complexity and platform is depended in the selection of wherein maximum level K, and to guarantee network, the span of K is between 3 ~ 8.
Obtaining these exhaustive guarantee relations to afterwards, just can according to certain rule, call statistical analysis unit 31 to these relations to classifying, after classification, the relation between each client is included into the class (step S6) in four kinds of relations (simple unidirectional guarantee one to one, one to one two-way mutual guarantor, three above colonies of client protect mutually, assure chain).Concrete, the basis for estimation following (supposing that two customer names are S and D) of four kinds of relations: 1) only there is unidirectional guarantee relation between client S and client D, this criterion directly can be calculated by statistical analysis unit 31, degree (client S)+degree (client D)=2.2) only there is two-way guarantee relation between two between client S and client D, this criterion directly can be calculated by statistical analysis unit 31 (such as in Examples below table 2 client A to client B and client B to client A all deposit guarantee and maximum level time is 1, be then simple two-way guarantee relation).3) if in the client of more than 3, all with other there is mutual guarantor in one or more user to any one user, be then mutual Bao Quan.4) if guarantee path does not exist circulation, and level time to be more than or equal to 3 (length (guarantee path) >=3 and there is not circulation), be then guarantee chain relation.The node client being included into the guarantee relation of this four classes relation is also included into identical classification.
The client node of corresponding four classes, needs to carry out a point cluster analysis, and point cluster analysis is exactly there is the client node connected each other to be labeled as same group, and hive off and have a lot of method, the complexity according to network structure is determined (step S7).Hive off according to being: simple unidirectional guarantee and two-way mutual guarantor one to one can be divided into colony respectively one to one; Three above colonies of client protect relation mutually and assure that chain relation clustering algorithm realizes, and realize with Modularity function 325.Wherein, Modularity algorithm needs parameters, determines that the principle of parameter needs the hardware configuration according to system deployment, network size, computational accuracy to require to determine, preferably
kbetween 10 ~ 100,
θbetween 0.1 ~ 0.001, more preferably
k=30,
θ=0.01, (iterations that k-is maximum,
θ-least error) can revise depending on data cases.
After obtaining these colonies, the function just can calling statistical analysis unit 31 carries out statistical study to these colonies, adds up the various group indexs such as client's number of each group, amount of guarantee, the promise breaking amount of money, rate of violation.And analysis is compared to the distinct group of similar index, different classes of feature is analyzed.By the statistics of these group indexs, in conjunction with the focus of business, can select the customer group paid close attention to, the standard of selection can carry out arranging (step S8) in statistical analysis software.
The customers' (step S8 generation) needing to pay close attention to are associated with the client node (step S5 generation) that needs are paid close attention to, every analytical applications can be met.Scheduling module 4 dispatches the every terms of information that query unit 11 can inquire about paid close attention to colony and client; Scheduling module 4 dispatch report unit 12 can represent the information such as pay close attention to colony and client every gathers, compares, change in the mode of form; Scheduling module 4 dispatching visualization unit 13 graphically can represent the incidence relation, coupling index, Changing Pattern etc. of paying close attention to colony and client; Required miscellaneous service conclusion can be obtained, the analysis (step S9) of step refining of going forward side by side, position subsequent by above analysis.
By above process, two kinds of results can be obtained:
1) view of each client, comprise other operational indicators that the centrad evaluation index that generated by assembly Degree function 321, Eigenvector Centrality function 322, Pagerank function 323 and statistical analysis unit 31 generate, constitute index view customer-centric.
2) classification of each colony and corresponding statistical indicator, comprises colony's numbering of four large classes, and the colony's level index system calculated by statistical analysis unit 31.
Based on above result, query unit 11 can provide the Flexible Query of client's level and colony's level, and form unit 12 can provide the report form showing of client's level and colony's level, visual function unit 14 can represent totally, locally, the characteristic sum relation of each colony.
Scheduling module 4 manually can call corresponding module in whole flow process, also can robotization call the analysis that corresponding module completes whole flow process, the data of generation and result are left in Data import and storage platform 1, for business diagnosis.
Illustrate that the present invention assures the work of the large data analysis system of circle below.
In step 2, Data import can load relevant business datum by the data model of its storage inside with memory module 2, such as, loaded with traffic data (table 1 is only citing data) shown in table 1.
Table 1
In the present embodiment, Data import and memory module 2 load data from external data, data comprise guarantee relation data, contract of guaranty data, customer default data etc., Data import and memory module 2 adopt JDBC mode to read from the mode of outside reading database, and the mode that the file layouts such as Excel then read parsing automatically from ftp is carried out.
After reading data, be exactly call indegree, outdegree index that Degree function 321 calculates each client respectively; Call Eigenvector Centrality function 322 calculate each client eigenvector centrality index, call the pagerank index that PageRank function 323 calculates each client.
After calculating these indexs, statistical analysis unit 31 pairs of index results are added up, and obtain the eigenvector centrality client of TOP 1000, indegree and the outdegree client of TOP 10%, the pagerank client etc. of TOP 50.Query unit 11 provides query function, and be the client etc. of TOP 10% as obtained 3 indexs, form unit 12 can associate all data, demonstrating data form, as the violation of agreement, key contracts situation etc. of Very Important Person simultaneously.
Ntree function 324 is piths for native system, and the data in certain level can be resolved by Ntree function.Ntree is the function based on java and MapReduce exploitation, can enumerate the relation between any one client having one or more levels guarantee contact, its computation process process flow diagram as shown in Figure 3.
Through ntree, the paired data that client assures relation can be obtained, such as, shown in table 2 (table 2 is only citing data).
Table 2
Obtain client pairing hierarchical relationship after, according to step S6 in Fig. 2 above-mentioned guarantee path is classified as simple unidirectional guarantee one to one, one to one two-way mutual guarantor by criteria for classification, three above colonies of client protect mutually, assure chain four class.
Judging to assure on the basis of relationship type, the client node in guarantee network can classified, be divided into above Four types.Wherein unidirectional guarantee, simple two-way guarantee and other classifications do not have co-user, and mutual Bao Quan and guarantee chain may exist common user.
After obtaining the type of each client node, hive off with regard to needs to every class, namely carry out cluster to the guarantee path generated by ntree, the method for cluster is as described in the step S7 of Fig. 2.After birdsing of the same feather flock together, obtain the clustering relationships between client and colony, such as, client shown in table 3 and the relation between colony (table 3 is only data of illustrating).
Table 3
Which colony client is arranged in as can be seen from Table 3, and as client 1, client 4, client 7 and client 8 all belong to colony 1, client 9, client 41 and client 85 all belong to colony 2; And the type of colony at client place can also be found out, if the colony at client 1, client 4, client 7 and client 8 place is for protect chain mutually, the colony at client 9, client 41 and client 85 place is mutual Bao Quan.
By above process, can obtain with the classification of the index diagram of each customers as center and each colony and corresponding statistical indicator, the colony's numbering comprising four large classes and the colony's level index system calculated by statistical analysis unit.
Claims (28)
1. for the analytic system that financial guaranty insurance is analyzed, comprising: scheduling module, analysis module and service display module;
Described scheduling module is for the work of the instruction coordinating analysis module and service display module that receive user;
Described analysis module is used for guarantee data analysis, and at least draws the index relevant to assuring the centrad of node;
Described service display module is used for the analysis result of analysis module to show with certain forms.
2. analytic system according to claim 1, is characterized in that: it is one or more that described service display module comprises in following functions unit:
Query unit, for inquiring about and specifying the guarantee information assured node or specify colony relevant;
Form unit, assures node for providing by the needs of user or specifies the form that colony is relevant to specifying;
Visualization, for realizing whole guarantee network and assuring the visual of colony's localized network.
3. analytic system according to claim 1, also comprises:
Data memory module, for store the raw data of input system, analysis module and service display CMOS macro cell process after data.
4. the analytic system according to any one of claims 1 to 3, is characterized in that:
It is described that to comprise to the relevant index of centrad of guarantee node in the out-degree of assuring node and in-degree index, centrality degree index of social network of overall importance, locality centrality degree index of social network one or more.
5. analytic system according to claim 4, is characterized in that:
Described out-degree and in-degree index are calculated by Degree function.
6. analytic system according to claim 4, is characterized in that:
Described centrality degree index of social network of overall importance is calculated by Eigenvector Centrality function.
7. analytic system according to claim 4, is characterized in that:
Described locality centrality degree index of social network is calculated by Pagerank function.
8. analytic system according to claim 4, is characterized in that:
Described analysis module carries out level calculating by enumerative technique to guarantee hoop net network, enumerates from a node to all guarantee path of another one node in guarantee network, and enumerates guarantee incidence relations all between any two nodes.
9. analytic system according to claim 8, is characterized in that:
When described analysis module carries out level calculating, the basis for selecting of start node is centrality degree index of social network of overall importance or the locality centrality degree index of social network of node, or random selecting.
10. analytic system according to claim 9, is characterized in that:
The maximum level scope that described analysis module carries out when level calculates is 3 ~ 8.
11. analytic systems according to claim 8, is characterized in that:
Guarantee relation is categorized into several basic form by the result that described analysis module calculates according to level.
12. analytic systems according to claim 11, is characterized in that:
Described several basic form comprises man-to-man unidirectional guarantee relation, man-to-man mutual guarantor's relation, three above colonies of client protect in relation, guarantee chain relation mutually one or more.
13. analytic systems according to claim 11, is characterized in that:
The result that level calculates by described analysis module is processed by Modularity function, thus guarantee relation is categorized into several basic form.
14. analytic systems according to claim 13, is characterized in that:
Described Modularity function uses Modularity Maximization method.
The analytical approach of 15. 1 kinds of financial guaranty insurances, comprising:
Scheduling step, receives the instruction of user and coordinates the work of service display module and analysis module;
Analytical procedure, to guarantee data analysis, and at least draws the index relevant to assuring the centrad of node;
Result shows step, is shown by analysis result with certain forms.
16. analytical approachs according to claim 15, it is characterized in that described result show step comprise the following steps in one or more:
Query steps, inquires about and feeds back and specify the guarantee information assured node or specify colony relevant;
Report generation step, provides by the needs of user and assures node or specify the form that colony is relevant to specifying;
Visualization processing step, carries out visual presentation to whole guarantee network and guarantee colony localized network.
17. analytical approachs according to claim 16, also comprise:
Data storing steps, for the data after stores processor.
18. analytical approachs according to any one of claim 15 ~ 17, is characterized in that:
It is described that to comprise to the relevant index of centrad of guarantee node in the out-degree of assuring node and in-degree index, centrality degree index of social network of overall importance, locality centrality degree index of social network one or more.
19. analytical approachs according to claim 18, is characterized in that:
Described out-degree and in-degree index are calculated by Degree function.
20. analytical approachs according to claim 18, is characterized in that:
Described centrality degree index of social network of overall importance is calculated by Eigenvector Centrality function.
21. analytical approachs according to claim 18, is characterized in that:
Described locality centrality degree index of social network is calculated by Pagerank function.
22. analytical approachs according to claim 18, is characterized in that:
By enumerative technique, level calculating is carried out to guarantee hoop net network, enumerate from a node to all guarantee path of another one node in guarantee network, and enumerate guarantee incidence relations all between any two nodes.
23. analytical approachs according to claim 22, is characterized in that:
The basis for selecting carrying out the start node of level calculating is centrality degree index of social network of overall importance or the locality centrality degree index of social network of node, or random selecting.
24. analytical approachs according to claim 23, is characterized in that:
The maximum level scope of carrying out level calculating is 3 ~ 8.
25. analytical approachs according to claim 22, is characterized in that:
According to the result that level calculates, guarantee relation is categorized into several basic form.
26. analytic systems according to claim 25, is characterized in that:
Described several basic form comprises man-to-man unidirectional guarantee relation, man-to-man mutual guarantor's relation, three above colonies of client protect in relation, guarantee chain relation mutually one or more.
27. analytical approachs according to claim 26, is characterized in that:
The result that level calculates is processed by Modularity function, thus guarantee relation is categorized into several basic form.
28. analytical approachs according to claim 27, is characterized in that:
Described Modularity function uses Modularity Maximization method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510342957.9A CN104933621A (en) | 2015-06-19 | 2015-06-19 | Big data analysis system and method for guarantee ring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510342957.9A CN104933621A (en) | 2015-06-19 | 2015-06-19 | Big data analysis system and method for guarantee ring |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104933621A true CN104933621A (en) | 2015-09-23 |
Family
ID=54120777
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510342957.9A Pending CN104933621A (en) | 2015-06-19 | 2015-06-19 | Big data analysis system and method for guarantee ring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104933621A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107767258A (en) * | 2017-09-29 | 2018-03-06 | 新华三大数据技术有限公司 | Risk of Communication determines method and device |
| CN109992435A (en) * | 2019-02-26 | 2019-07-09 | 上海拍拍贷金融信息服务有限公司 | Equipment uses abnormality determination method, device and computer storage medium |
| CN111754340A (en) * | 2020-07-03 | 2020-10-09 | 交通银行股份有限公司 | Guarantee network risk investigation system based on graph database |
| CN111784528A (en) * | 2020-05-27 | 2020-10-16 | 平安科技（深圳）有限公司 | Abnormal community detection method and device, computer equipment and storage medium |
| CN111784495A (en) * | 2020-06-04 | 2020-10-16 | 江苏常熟农村商业银行股份有限公司 | Guarantee ring identification method and device, computer equipment and storage medium |
| CN113780694A (en) * | 2020-06-10 | 2021-12-10 | 阿里巴巴集团控股有限公司 | Risk assessment method and device and electronic equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101661482A (en) * | 2008-08-27 | 2010-03-03 | 国际商业机器公司 | Method and device for recognizing similar subgraph in network |
| CN102609817A (en) * | 2012-02-14 | 2012-07-25 | 北京久其软件股份有限公司 | Display method and display device for enterprise data |
| CN102903056A (en) * | 2011-07-27 | 2013-01-30 | 刘金衢 | Method and system for debit and credit guarantee based on e-commerce platform |
| CN103020302A (en) * | 2012-12-31 | 2013-04-03 | 中国科学院自动化研究所 | Academic core author excavation and related information extraction method and system based on complex network |
| CN103218400A (en) * | 2013-03-15 | 2013-07-24 | 北京工业大学 | Method for dividing network community user groups based on link and text contents |
| CN103412878A (en) * | 2013-07-16 | 2013-11-27 | 西安交通大学 | Document theme partitioning method based on domain knowledge map community structure |
| JP2014021862A (en) * | 2012-07-20 | 2014-02-03 | Nippon Telegr & Teleph Corp <Ntt> | Associated node searching device, associated node searching method and program |
| CN104346367A (en) * | 2013-07-30 | 2015-02-11 | 华为技术有限公司 | Data integration method and device |
-
2015
- 2015-06-19 CN CN201510342957.9A patent/CN104933621A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101661482A (en) * | 2008-08-27 | 2010-03-03 | 国际商业机器公司 | Method and device for recognizing similar subgraph in network |
| CN102903056A (en) * | 2011-07-27 | 2013-01-30 | 刘金衢 | Method and system for debit and credit guarantee based on e-commerce platform |
| CN102609817A (en) * | 2012-02-14 | 2012-07-25 | 北京久其软件股份有限公司 | Display method and display device for enterprise data |
| JP2014021862A (en) * | 2012-07-20 | 2014-02-03 | Nippon Telegr & Teleph Corp <Ntt> | Associated node searching device, associated node searching method and program |
| CN103020302A (en) * | 2012-12-31 | 2013-04-03 | 中国科学院自动化研究所 | Academic core author excavation and related information extraction method and system based on complex network |
| CN103218400A (en) * | 2013-03-15 | 2013-07-24 | 北京工业大学 | Method for dividing network community user groups based on link and text contents |
| CN103412878A (en) * | 2013-07-16 | 2013-11-27 | 西安交通大学 | Document theme partitioning method based on domain knowledge map community structure |
| CN104346367A (en) * | 2013-07-30 | 2015-02-11 | 华为技术有限公司 | Data integration method and device |
Non-Patent Citations (6)
| Title |
|---|
| 刘东庆: "信贷担保圈的识别与管理", 《全国商情(理论研究)》 * |
| 刘军著: "《整体网分析讲义 UCINET软件实用指南》", 31 July 2009, 格致出版社 * |
| 吴晓秋等: "基于关键词共现频率的热点分析方法研究", 《情报理论与实践》 * |
| 威海银监分局担保圈风险研究课题组: "基于有向图遍历算法的担保圈风险监测预警系统研究——从商业银行的视角出发", 《金融监管研究》 * |
| 美）唐磊等著: "《社会计算 社区发现和社会媒体挖掘》", 31 January 2013, 机械工业出版社 * |
| 郑杲娉等: "影子银行网络与资本市场的系统性风险_基于沪深担保圈的经验证据", 《成都大学学报》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107767258A (en) * | 2017-09-29 | 2018-03-06 | 新华三大数据技术有限公司 | Risk of Communication determines method and device |
| CN107767258B (en) * | 2017-09-29 | 2021-07-02 | 新华三大数据技术有限公司 | Risk propagation determination method and device |
| CN109992435A (en) * | 2019-02-26 | 2019-07-09 | 上海拍拍贷金融信息服务有限公司 | Equipment uses abnormality determination method, device and computer storage medium |
| CN111784528A (en) * | 2020-05-27 | 2020-10-16 | 平安科技（深圳）有限公司 | Abnormal community detection method and device, computer equipment and storage medium |
| CN111784495A (en) * | 2020-06-04 | 2020-10-16 | 江苏常熟农村商业银行股份有限公司 | Guarantee ring identification method and device, computer equipment and storage medium |
| CN111784495B (en) * | 2020-06-04 | 2021-05-04 | 江苏常熟农村商业银行股份有限公司 | Guarantee ring identification method and device, computer equipment and storage medium |
| CN113780694A (en) * | 2020-06-10 | 2021-12-10 | 阿里巴巴集团控股有限公司 | Risk assessment method and device and electronic equipment |
| CN111754340A (en) * | 2020-07-03 | 2020-10-09 | 交通银行股份有限公司 | Guarantee network risk investigation system based on graph database |
| CN111754340B (en) * | 2020-07-03 | 2022-06-21 | 交通银行股份有限公司 | Guarantee network risk investigation system based on graph database |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kamble et al. | Modeling the blockchain enabled traceability in agriculture supply chain | |
| Celik et al. | A comprehensive review of multi criteria decision making approaches based on interval type-2 fuzzy sets | |
| CN104933621A (en) | Big data analysis system and method for guarantee ring | |
| Danesh et al. | A systematic comparison of multi-criteria decision making methods for the improvement of project portfolio management in complex organisations | |
| US8340995B2 (en) | Method and system of using artifacts to identify elements of a component business model | |
| Zahradníčková et al. | Scenarios as a strong support for strategic planning | |
| Mohammed et al. | A hybrid MCDM approach towards resilient sourcing | |
| CN106067094A (en) | A kind of dynamic assessment method and system | |
| Pandey et al. | Supply chain risks in Industry 4.0 environment: review and analysis framework | |
| CN103336790A (en) | Hadoop-based fast neighborhood rough set attribute reduction method | |
| Meng et al. | Decision making with intuitionistic linguistic preference relations | |
| Salahi et al. | Robust Russell and enhanced Russell measures in DEA | |
| Hsu et al. | Deploying big data enablers to strengthen supply chain resilience to mitigate sustainable risks based on integrated HOQ-MCDM framework | |
| CN103336791A (en) | Hadoop-based fast rough set attribute reduction method | |
| Xing et al. | Betweenness centrality for similarity-weight network and its application to measuring industrial sectors’ pivotability on the global value chain | |
| Jallan et al. | Text mining of the securities and exchange commission financial filings of publicly traded construction firms using deep learning to identify and assess risk | |
| Sun et al. | Data mining for evaluating the ecological compensation, static and dynamic benefits of returning farmland to forest | |
| Adeniran et al. | The significance of strategic management accounting on the performance of transport businesses in Nigeria | |
| CN109102176A (en) | A kind of trade finance appraisal procedure and device based on block chain | |
| Podvezko et al. | Assessing the performance of the construction sectors in the Baltic states and Poland | |
| Bryan | The assessment of social impacts | |
| Renigier-Biłozor et al. | Optimization of the variables selection in the process of real estate markets rating | |
| Putri et al. | Big data and strengthening msmes after the covid-19 pandemic (development studies on batik msmes in east java) | |
| Li et al. | Predicting business risks of commercial banks based on BP-GA optimized model | |
| Hashemkhani Zolfani et al. | Application of MCDM methods in short-term planning for private universities based on balanced scorecard: a case study from Iran |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150923 |
|
| WD01 | Invention patent application deemed withdrawn after publication |