CN101464877B - Method for digging related criminal suspect - Google Patents
Method for digging related criminal suspect Download PDFInfo
- Publication number
- CN101464877B CN101464877B CN 200810121668 CN200810121668A CN101464877B CN 101464877 B CN101464877 B CN 101464877B CN 200810121668 CN200810121668 CN 200810121668 CN 200810121668 A CN200810121668 A CN 200810121668A CN 101464877 B CN101464877 B CN 101464877B
- Authority
- CN
- China
- Prior art keywords
- criminal
- node
- suspect
- crime
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000005314 correlation function Methods 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 12
- 238000013139 quantization Methods 0.000 claims description 8
- 230000010354 integration Effects 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000003012 network analysis Methods 0.000 abstract description 4
- 230000002596 correlated effect Effects 0.000 abstract 2
- 230000000875 corresponding effect Effects 0.000 abstract 1
- 238000011835 investigation Methods 0.000 abstract 1
- 238000009412 basement excavation Methods 0.000 description 4
- 238000007418 data mining Methods 0.000 description 4
- 238000013459 approach Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000001568 sexual effect Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000010205 computational analysis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000284 resting effect Effects 0.000 description 1
Landscapes
- Alarm Systems (AREA)
Abstract
The invention relates to the field of network analysis for crime in public security, in particular to a method thereof for identifying relevant criminal suspects. The invention further provides a method for identifying relevant criminal suspects: 1), finding a corresponding node of a main criminal suspect in a criminal network; 2), constructing a collection of correlative personnel; 3), expanding the correlation according to cross-correlation of personnel;4), adding the expanded nodes with a criminal correlation coefficient value higher than a threshold value into the collection of the correlative personnel; 5), repeating step 3) and step 4) until no new node is added into the collection of correlative personnel or all nodes of the criminal network are expanded; and 6), selecting a quantitative node as a correlated suspect for investigation according to the ordering of the criminal correlation coefficient between each nodes and the main criminal suspect. The identifying method thereof have the advantages that the efficiency is greatly improved, and the correlated criminals can be fully screened out.
Description
Technical field
The present invention relates to public security criminal network analysis field, refer to especially the method for digging related criminal suspect in a kind of extensive criminal network.
Background technology
Since 911 events, the security department of every country attaches great importance to the collection of crime data and the construction of Relational database, is carrying out a series of research aspect the data mining based on Relational database simultaneously.The analysis of crime data and excavation are to start with from dominant data, seek out some recessive useful information.The psychology that may hide from criminal network, behavior or other factor extract relevant pattern, give a clue, assist to handle a case with this, simultaneously, among as much as possible the feature mode that extracts being used for monitoring, realize early warning.
Data mining (Data Mining) is intended to from data a large amount of, incomplete, noisy, fuzzy, at random, extract lie in wherein, people are ignorant but be information and the knowledge of potentially useful in advance.Also have a lot of and the akin term of this term, as from database, finding knowledge (KDD), data analysis, data fusion (Data Fusion) and decision support etc.Mention in the same breath in these concepts, be because data mining aiming to as if set forth the correlativity knowledge that contacts between individuality but not describe individual attribute, zero broken data.
Existing criminal network analytical approach is from manual method to the at last up till now social network analysis of comparative maturity (SNA) of analysis based on picture showing.Social network analysis has been widely used in analysis, the analysis of social character relation, the discovery of social structure and the research of understanding and organizing communication behavior of international trade relation at present.But all analyses also only rest on the aspect of Manual analysis at present, namely remain the criminal network analytical approach of the so-called first generation.When this analytical approach was faced huge data volume, efficient was not very high, and has certain randomness.The policeman normally relies on for many years experience on purpose to be sought oneself to think Useful Information in boundless and indistinct data, in fact the analysis of data still is to finish by manpower fully
Some too theorizes existing digging system and method, perhaps only has method and lacks the accumulation of a given data; Some can be practical, high performance method only can process the criminal network of middle and small scale; And some is for the tool and method of extensive criminal network, the graphical displaying after only resting on preliminary excavation or laying particular emphasis on artificial treatment, the role who only takes on auxiliary type aspect excavation.These digging systems and method for digging efficient are not high, rely on manpower analysis, and randomness is very strong, omit easily.
Summary of the invention
The object of the present invention is to provide a kind of efficient in extensive criminal network, comprehensive method of digging related criminal suspect.
A kind of method of digging related criminal suspect, the method comprises the steps:
1) user inputs center suspect's information, finds corresponding node in criminal network;
2) utilize crime correlation function module structure associate people set, the center suspect's that its finite element is determined for the user node;
3) according to the personnel's cross-correlation in the criminal network, choose each node that was not expanded in the associate people set, be located in criminal network, and in this network, carry out the association expansion;
4) utilize crime correlation function module to calculate the crime correlation coefficient value that is expanded node, this value is not less than the node join associate people set of threshold value;
5) repeating step 3) and 4), until the associate people set no longer includes new node join or the whole nodes of criminal network all are expanded complete;
6) according to each node in the associate people set and center suspect's crime correlation coefficient, utilize related quantization modules to sort, get a certain amount of node and investigate as candidate's the suspect that is associated,
Wherein,
Described step 1) criminal network is related optimization graph theory model between the possible crime individuality of quantificational expression in, and model tormulation is as follows:
G=(c, C, A, S, f), wherein:
G: criminal network,
C: the center suspect, by user's input,
C: all set of node in the criminal network,
A: all related set of crime in the criminal network,
F: calculate the function of crime correlation coefficient,
Described step 2) and step 4) in the correlation function f of crime correlation function module be defined as:
N: the element number of set C;
C: the center suspect, inputted by the user;
A
Ij: node i, the mainovre relating value between the j,
Described step 6) related quantization modules utilizes the Analysis of Hierarchy Structure method to determine that each is associated in the priority in the network in; Priority being mapped to fifty-fifty [0,1] space quantizes.
Further, described step 4) the newly-increased node of associate people set is a part of crowd who is associated with the center suspect in the criminal network in, is defined as: S={x | f (x, c) 〉=k}, and wherein k is the crime correlation threshold; Initial situation S={c}, c: the center suspect, inputted by the user.
Further, described step 6) comprise that also utilizing in case of necessity node to screen module makes personnel's relation integration size be controlled at all the time the interior step of scope that manpower can be investigated.
The present invention compares with background technology, and the useful effect that has is: realize that node quantity is millions (10
7~10
8) criminal network scan fully; The growth of control association personnel set is so that this scanning process can be finished on computing power is subject to the database server of larger restriction, low side quickly; The crime association is screened all sidedly, adopts again the AHP method to determine priority, use at last average geometric ratio to quantize, fully taken into account miscellaneous category of offenses be associated in handle a case, the complicated variety in the computational analysis so that criminal network model approaching to reality situation more; The good effectively crime correlation function of design has deeply excavated and has had relevance between the criminal offence, and so that can restrain in the more shallow degree of depth in the operation of the great majority on the huge criminal network, remains the locality access; And adopting the method for digging of determining, is a kind of stable, reproducible method.
The present invention be directed to the different case backgrounds of identical network and be optimized modeling, the refinement extraction and analysis is carried out in the miscellaneous category of offenses association, and then launch on this basis optimal control candidate nodal point number and measure suspect's Result.The present invention is by setting up the criminal network that inclusion information is abundanter, relevance is stronger, add suitable quantization method, good crime correlation function structure and, thereby guaranteed the faster accuracy rate of excavation speed and Geng Gao.
Description of drawings
Fig. 1 is criminal network and associate people set.
Embodiment
The system of digging related criminal suspect mainly comprises related quantization modules, crime correlation function module and node screening module among the present invention.
1, related quantization modules
This module is responsible for determining of associated priority and is quantized.At first integrate a plurality of public business databases, social personnel's related information and offender's related information are merged, and construct a criminal network; Various related informations in this network between the extraction personnel, such as: the accomplice is related, association is had sexual intercourse in the hotel, telephone relation is related, the aircraft colleague is related, relatives are related, the colleague is related, the fellow villager is related etc.; Adopt existing mature technology Analysis of Hierarchy Structure method (AHP) to determine that each is associated in the priority in the network; Priority is mapped to fifty-fifty [0,1] space and then finishes quantizing process; Associated priority can change according to the difference of case, can embody the correlation difference opposite sex by revising judgment matrix, so quantized result can change also.
R is the associative expression formula, and there is relation integration r in R (x, y, r) expression node x between the y; Wherein
Ur={ accomplice, have sexual intercourse in the hotel ..., be related complete or collected works;
The priority of the related rx of Pr (rx) expression, wherein rx ∈ Ur; Pr (hotel the same branch of a family)>Pr (fellow villager) for example can be set;
The quantized values of the related rx of Q (rx) expression increases with the increase of priority.
2, crime correlation function module
This module is responsible for the calculating of crime correlation coefficient.When calculating the crime correlation coefficient of i node, judge that at first whether i exists the mainovre relation with center suspect c, temporarily is made as Pr (ric) with f (i, c) value first if exist, otherwise is made as 0; From associate people set, choose the node that does not compare again, judge that whether node i carries out related coefficient greater than f (i, c) by all the other any node j ∈ S with c, if greater than just upgrade f (, c) being worth is Pr (rij) * f (j, c); Thereby can tentatively weigh the correlation degree of i node and center suspect c.
The definition of correlation function f:
N: the element number of set C;
Aij: node i, the mainovre relating value between the j;
Be the set of criminal network and associate people such as Fig. 1, the numeral in the bracket this with the crime correlation coefficient of center suspect c, be provided with k=0.35, so C
1, C
2, C
4, C
6Four nodes and c have consisted of the associate people set jointly.
3, node screening module
This module is responsible for maintaining relation integration, and its size is controlled in the scope that manpower can investigate all the time.When each node calculates when complete, if the quantity of relation integration has exceeded the maximum nodal point number of default set, so each element of relation integration crime correlation coefficient according to them is sorted, begin deletion from the node with minimum crime correlation coefficient, until not deleted node quantity is no more than the maximum nodal point number of set; If the nodal point number of relation integration does not surpass the maximum nodal point number of set, with the node deletion of correlation coefficient less than threshold value k.
The method of digging related criminal suspect of the present invention comprises the steps:
1) user inputs center suspect's ID (identity number) card No., finds corresponding node in criminal network;
2) utilize the set of crime correlation function module structure associate people, the center suspect that its finite element is determined for the user;
3) according to the personnel's cross-correlation in the criminal network, choose each node that was not expanded in the associate people set, be located in criminal network, and in this network, carry out the association expansion;
4) utilize crime correlation function module to calculate the crime correlation coefficient value that is expanded node, this value is not less than the node join associate people set of threshold value;
5) repeating step (3) and (4) are until the associate people set no longer includes new node join or the whole nodes of criminal network all are expanded complete.
6) utilize related quantization modules, sort according to each node in the associate people set and center suspect's crime correlation coefficient, get a certain amount of node and investigate as candidate's the suspect that is associated.
Wherein,
Step 1) criminal network of determining in is related optimization graph theory model between the possible crime individuality of quantificational expression, and model tormulation is as follows:
G: criminal network, G=(c, C, A, S, f), wherein:
C: the center suspect, inputted by the user;
C: all set of node in the criminal network;
A: all related set of crime in the criminal network;
F: the function that calculates the crime correlation coefficient;
Step 4) the associate people set in: the node in this set has represented a part of crowd who is associated with the center suspect in the criminal network, is defined as: S={x | f (x, c) 〉=k}, and wherein k is the crime correlation threshold; Initial situation S={c};
Step 4) the crime correlation function in: f is for calculating the function of crime correlation coefficient.F (i, c) is expressed as the crime correlation degree of node i and center node c.Formal definitions is:
N: the element number of set C;
A
Ij: node i, the mainovre relating value between the j;
Claims (3)
1. the method for a digging related criminal suspect, it is characterized in that: the method comprises the steps:
1) user inputs center suspect's information, finds corresponding node in criminal network;
2) utilize crime correlation function module structure associate people set, the center suspect's that its finite element is determined for the user node;
3) according to the personnel's cross-correlation in the criminal network, choose each node that was not expanded in the associate people set, be located in criminal network, and in this network, carry out the association expansion;
4) utilize crime correlation function module to calculate the crime correlation coefficient value that is expanded node, this value is not less than the node join associate people set of threshold value;
5) repeating step 3) and 4), until the associate people set no longer includes new node join or the whole nodes of criminal network all are expanded complete;
6) according to each node in the associate people set and center suspect's crime correlation coefficient, utilize related quantization modules to sort, get a certain amount of node and investigate as candidate's the suspect that is associated,
Wherein,
Described step 1) criminal network is related optimization graph theory model between the possible crime individuality of quantificational expression in, and model tormulation is as follows:
G=(c, C, A, S, f), wherein:
G: criminal network,
C: the center suspect, by user's input,
C: all set of node in the criminal network,
A: all related set of crime in the criminal network,
F: calculate the function of crime correlation coefficient,
Described step 2) and step 4) in the correlation function f of crime correlation function module be defined as:
N: the element number of set C;
C: the center suspect, inputted by the user;
Aij: node i, the mainovre relating value between the j,
Described step 6) related quantization modules utilizes the Analysis of Hierarchy Structure method to determine that each is associated in network in
In priority; Priority being mapped to fifty-fifty [0,1] space quantizes.
2. the method for described digging related criminal suspect according to claim 1, it is characterized in that: the newly-increased node of associate people set is a part of crowd who is associated with the center suspect in the criminal network described step 4), be defined as: S={x | f (x, c) 〉=and k}, wherein k is the crime correlation threshold; Initial situation S={c}, c: the center suspect, inputted by the user.
3. the method for described digging related criminal suspect according to claim 1 is characterized in that: described step 6) also comprise and utilize in case of necessity node screening module to make personnel's relation integration size be controlled at all the time step in the scope that manpower can investigate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200810121668 CN101464877B (en) | 2008-10-27 | 2008-10-27 | Method for digging related criminal suspect |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200810121668 CN101464877B (en) | 2008-10-27 | 2008-10-27 | Method for digging related criminal suspect |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101464877A CN101464877A (en) | 2009-06-24 |
CN101464877B true CN101464877B (en) | 2013-01-30 |
Family
ID=40805457
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 200810121668 Expired - Fee Related CN101464877B (en) | 2008-10-27 | 2008-10-27 | Method for digging related criminal suspect |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101464877B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106610997A (en) * | 2015-10-23 | 2017-05-03 | 杭州海康威视数字技术股份有限公司 | Method, device and system for processing person information |
Families Citing this family (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105138570B (en) * | 2015-07-26 | 2019-02-05 | 吉林大学 | The doubtful crime degree calculation method of network speech data |
CN106127231A (en) * | 2016-06-16 | 2016-11-16 | 中国人民解放军国防科学技术大学 | A kind of crime individual discrimination method based on the information Internet |
CN106682990A (en) * | 2016-12-09 | 2017-05-17 | 武汉中软通证信息技术有限公司 | Method and system for establishing interpersonal relationship model of suspect |
CN109325814A (en) * | 2017-07-31 | 2019-02-12 | 上海诺悦智能科技有限公司 | A method of for finding suspicious trade network |
CN107679201B (en) * | 2017-10-12 | 2018-08-31 | 杭州中奥科技有限公司 | Hide people's method for digging, device and electronic equipment |
CN108052641A (en) * | 2017-12-21 | 2018-05-18 | 中通服公众信息产业股份有限公司 | The personnel's infectiosity coefficient calculation method calculated based on large scale network |
CN109857557A (en) * | 2019-01-16 | 2019-06-07 | 北京明略软件系统有限公司 | The distributed computing method and system, computer readable storage medium of relationship discovery |
CN109947870A (en) * | 2019-03-26 | 2019-06-28 | 第四范式(北京)技术有限公司 | The prediction meanss and method of specific type personnel calculate equipment and storage medium |
CN111177192A (en) * | 2019-12-11 | 2020-05-19 | 北京明略软件系统有限公司 | Method and device for determining group members |
-
2008
- 2008-10-27 CN CN 200810121668 patent/CN101464877B/en not_active Expired - Fee Related
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106610997A (en) * | 2015-10-23 | 2017-05-03 | 杭州海康威视数字技术股份有限公司 | Method, device and system for processing person information |
CN106610997B (en) * | 2015-10-23 | 2019-05-21 | 杭州海康威视数字技术股份有限公司 | People information processing method, apparatus and system |
Also Published As
Publication number | Publication date |
---|---|
CN101464877A (en) | 2009-06-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101464877B (en) | Method for digging related criminal suspect | |
US11580104B2 (en) | Method, apparatus, device, and storage medium for intention recommendation | |
CN103605665B (en) | Keyword based evaluation expert intelligent search and recommendation method | |
CN108052576A (en) | A kind of reason knowledge mapping construction method and system | |
CN111382276B (en) | Event development context graph generation method | |
CN108596386A (en) | A kind of prediction convict repeats the method and system of crime probability | |
CN104346355B (en) | The intelligent search method and its system of serial public security case | |
CN106778876A (en) | User classification method and system based on mobile subscriber track similitude | |
CN108509614A (en) | A kind of task record management and analysis method based on chart database | |
CN105224593B (en) | Frequent co-occurrence account method for digging in the of short duration online affairs of one kind | |
CN101655857A (en) | Method for mining data in construction regulation field based on associative regulation mining technology | |
CN106980651B (en) | Crawling seed list updating method and device based on knowledge graph | |
CN108647322A (en) | The method that word-based net identifies a large amount of Web text messages similarities | |
JP2002032773A (en) | Device and method for processing map data | |
CN106528608A (en) | Cold and hot storage method and system for power grid GIS (Geographic Information System) data in cloud architecture | |
CN104317794B (en) | Chinese Feature Words association mode method for digging and its system based on dynamic item weights | |
CN103761286B (en) | A kind of Service Source search method based on user interest | |
CN109597926A (en) | A kind of information acquisition method and system based on social media emergency event | |
CN102254025B (en) | Information memory retrieving method | |
CN106355537A (en) | Smart analysis method and system for interrelated cases | |
CN112989215A (en) | Knowledge graph enhanced recommendation system based on sparse user behavior data | |
CN106294688A (en) | Query expansion method, device and system based on user characteristic analysis | |
Nishino et al. | A location predictor based on dependencies between multiple lifelog data | |
CN110502521B (en) | Method for establishing archive | |
CN105447013A (en) | News recommendation system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130130 Termination date: 20171027 |
|
CF01 | Termination of patent right due to non-payment of annual fee |