CN106355537A - Smart analysis method and system for interrelated cases - Google Patents
Smart analysis method and system for interrelated cases Download PDFInfo
- Publication number
- CN106355537A CN106355537A CN201610711740.5A CN201610711740A CN106355537A CN 106355537 A CN106355537 A CN 106355537A CN 201610711740 A CN201610711740 A CN 201610711740A CN 106355537 A CN106355537 A CN 106355537A
- Authority
- CN
- China
- Prior art keywords
- data
- case
- track
- related cases
- analysis
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 153
- 238000010801 machine learning Methods 0.000 claims abstract description 44
- 230000000694 effects Effects 0.000 claims description 25
- 238000001514 detection method Methods 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 16
- 230000010354 integration Effects 0.000 claims description 14
- 238000012546 transfer Methods 0.000 claims description 11
- 238000007405 data analysis Methods 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 abstract description 8
- 238000011840 criminal investigation Methods 0.000 abstract description 3
- 238000000034 method Methods 0.000 description 38
- 239000000284 extract Substances 0.000 description 11
- 238000000605 extraction Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 10
- 239000002699 waste material Substances 0.000 description 6
- 238000011835 investigation Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000002411 adverse Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000013508 migration Methods 0.000 description 2
- 230000005012 migration Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000001568 sexual effect Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000010224 classification analysis Methods 0.000 description 1
- 238000000205 computational method Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000002650 habitual effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000013517 stratification Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/25—Integrating or interfacing systems involving database management systems
- G06F16/254—Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Economics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
This invention discloses a smart analysis method and system for interrelated cases. It falls within the technical field of public security case analysis, and solves the time-consuming and low-efficient issues that the current technology available requires investigators to artificially analyze and search for the needed information from the public security information system when it comes to the analysis of interrelated cases, thereby boosting the efficiency and accuracy in analyzing interrelated cases. Among them, the smart analysis method for interrelated cases includes the following steps: establish a case characteristic data warehouse based on the public security criminal investigation database, generate a analysis model through machine learning based on the said case characteristics data warehouse, use the said case analysis model for interrelated cases to analyze the unsolved cases to generate a list of assorted case record.
Description
Technical field
The present invention relates to public security case analysis technical field, in particular to a kind of intelligence combine related cases analysis method and
System.
Background technology
Currently, the situation very severe of criminal offense struggle, criminal case incidence of criminal offenses constantly rises, the notable spy of criminal activity
Point is fleeing property and serial case than more prominent.Offender is using trans-regional, saltatory, the method fled on a large scale is carried out
Crime, means of crime is increasingly cunning, and violence nature becomes apparent from, and space-time span is bigger, and intelligent, the poly-talented clique oriented depth of crime
Layer direction development trend.For fleeing property of crime, serial feature, part work of combining related cases should be strengthened, carry out combining related cases part
Investigation.
Analysis of combining related cases is an important investigation of criminal investigation work, and so-called analysis of combining related cases refers to by case
The crime means of part, vestige, material evidence etc. carry out comprehensive analysis, and the case that there will be certain contact combines to merge and detects
Broken methods of investigation.More case information can be obtained by combining related cases, to serial, seriality, multiple and professional
The detection of sexual crime case has great effect.Part of combining related cases is conducive to strengthening trans-regional criminal investigation cooperation, shared Crime Information
Resource, is conducive to obtaining various evidence of crime, deep-cuts remaining crime, crime prevention.The main method of part of combining related cases has: conspired with case,
Conspired with people, conspired with thing, conspired with case property.
Typically require investigator in traditional method manually to be combined related cases from conventional case data base analysis, so
And, due to being continuously increased of crime case, in police informatization, the case data volume of storage is extremely huge, additionally, some are no
The redundancy of effect also occurs, investigator wants to find by way of artificial from police informatization related with case
Case, degree of difficulty is well imagined, which results in the increase of manpower, the decline of efficiency.Based on original Police Information
The method of combining related cases of system, frequently can lead to final combining related cases and analyzes not comprehensively, and not prompt enough accurate, for example, should enter
String case simultaneously of going is not gone here and there and is gone here and there simultaneously simultaneously or in time, thus having affected adversely the opportunity of solving a case, serious also can cause the personnel of committing a crime continuous
Crime, becomes habitual offender, generates a panic to citizen and fear, and the harmony having a strong impact on society is stable.
Content of the invention
In view of this, analysis method and system it is an object of the invention to provide a kind of intelligence is combined related cases, existing to solve
Combine related cases in technology analysis when need investigator from police informatization manual analyses search lead to waste time and energy, efficiency
Low problem, it is possible to increase combine related cases analysis efficiency and accuracy rate.
In a first aspect, embodiments providing a kind of intelligent analysis method of combining related cases, methods described includes following step
Rapid:
Based on police criminal detection Database case characteristic warehouse;
Analysis model of combining related cases is generated by machine learning based on described case characteristic warehouse;
Using described analysis model of combining related cases, unsolved case is analyzed, generates class case list.
In conjunction with a first aspect, embodiments providing the first possible embodiment of first aspect, wherein, on
State method further comprising the steps of:
High-risk group data base is set up based on public security previous conviction demographic data storehouse.
In conjunction with a first aspect, the possible embodiment of the second that embodiments provides first aspect, wherein, on
State method further comprising the steps of:
Personnel activity's track database is set up based on trace information.
In conjunction with a first aspect, embodiments providing the third possible embodiment of first aspect, wherein, on
State method further comprising the steps of:
Transfer high-risk group's data from described high-risk group data base;
Transfer track data from described personnel activity's track database;
Using described high-risk group's data and described track data, class case list is carried out with data analysiss, generate row of combining related cases
Table.
In conjunction with a first aspect, embodiments providing the 4th kind of possible embodiment of first aspect, wherein,
Described trace information includes fence track, online track, lodging track, consumption track, transit trip
Track.
In conjunction with a first aspect, embodiments providing the 5th kind of possible embodiment of first aspect, wherein, institute
State machine learning and specifically include following steps:
Select training experience;
Selection target function;
The representation of selection target function;
Select Function Approximation Algorithm.
Embodiments provide a kind of intelligent analysis method of combining related cases, by setting up based on police criminal detection data base's
Case characteristic warehouse, supports as data.Machine learning is passed through based on case characteristic warehouse, analysis is broken to combine related cases
The feature of part, generate combine related cases analysis model it is possible to user actually used middle continuous renewal, perfect, finally using string
And case analysis model is analyzed to unsolved case, generate class case list.In Criminal characteristic, there is identical point, not the breaking of similitude
Case, will act as class case and occurs in such case list.Thus solve to need when combining related cases analysis in prior art to investigate
Personnel's manual analyses from police informatization search lead to waste time and energy, the problem of inefficiency, improve and combine related cases point
Analysis efficiency and accuracy rate, reduction missing rate.
Second aspect, embodiments provides a kind of intelligent analysis system of combining related cases, comprising:
Data integration module, for based on police criminal detection Database case characteristic warehouse;
Machine learning module, for generating, by machine learning, analysis mould of combining related cases based on described case characteristic warehouse
Type;
Class case generation module, for being analyzed to unsolved case using described analysis model of combining related cases, generates class case row
Table.
In conjunction with second aspect, inventive embodiments provide the first possible embodiment of second aspect, wherein, this intelligence
Analysis system of combining related cases also includes:
Trajectory analysis module, for setting up personnel activity's track database based on trace information;
High-risk group's generation module, for setting up high-risk group data base based on public security previous conviction demographic data storehouse.
In conjunction with second aspect, inventive embodiments provide the possible embodiment of second of second aspect, wherein, this intelligence
Analysis system of combining related cases also includes:
Data access module, for adjusting respectively from described personnel activity's track database, described high-risk group data base
Take personnel's track data, high-risk group's data;
Analysis module, divides for carrying out data using described high-risk group's data and described track data to class case list
Analysis, generates list of combining related cases.
In conjunction with second aspect, inventive embodiments provide the third possible embodiment of second aspect, wherein,
Described trace information includes fence track, online track, lodging track, consumption track, transit trip
Track.
Embodiments provide a kind of intelligent analysis system of combining related cases, including data integration module, machine learning mould
Block and class case generation module, wherein, data integration module is used for based on police criminal detection Database case characteristic warehouse,
Machine learning module is used for generating, by machine learning, analysis model of combining related cases based on case characteristic warehouse, and class case generates mould
Block is used for using combining related cases point
Analysis model is analyzed to unsolved case, generates class case list.By setting up case characteristic warehouse, as number
According to support, it is then based on case characteristic warehouse and passes through machine learning, the feature of the broken part of combining related cases of analysis, generation is combined related cases
Analysis model it is possible to user actually used middle continuous renewal, perfect, finally using analysis model of combining related cases to not solving a case
Part is analyzed, and generates class case list.There is in Criminal characteristic the unsolved case of identical point, similitude, will act as class case and go out
In such case list now.Thus solve needing investigator from police informatization when combining related cases analysis in prior art
Manual analyses search lead to waste time and energy, the problem of inefficiency, improve combine related cases analysis efficiency and accuracy rate.
Other features and advantages of the present invention will illustrate in the following description, and, partly become from description
Obtain it is clear that or being understood by implementing the present invention.The purpose of the present invention and other advantages are in description, claims
And in accompanying drawing specifically noted structure realizing and to obtain.
For enabling the above objects, features and advantages of the present invention to become apparent, preferred embodiment cited below particularly, and coordinate
Appended accompanying drawing, is described in detail below.
Brief description
In order to be illustrated more clearly that the technical scheme of the embodiment of the present invention, below will be attached to use required in embodiment
Figure is briefly described it will be appreciated that the following drawings illustrate only certain embodiments of the present invention, and it is right to be therefore not construed as
The restriction of scope, for those of ordinary skill in the art, on the premise of not paying creative work, can also be according to this
A little accompanying drawings obtain other related accompanying drawings.
Fig. 1 shows that a kind of intelligence that the embodiment of the present invention is provided is combined related cases the flow chart of analysis method;
Fig. 2 shows that a kind of intelligence that the embodiment of the present invention is provided is combined related cases the flow process of machine learning in analysis method
Figure;
Fig. 3 shows that a kind of intelligence that the embodiment of the present invention is provided is combined related cases the structural framing figure of analysis system.
Illustrate:
301- data integration module, 302- machine learning module, 303- class case generation module
304- trajectory analysis module, 305- high-risk group's generation module
306- data access module, 307- analysis module
Specific embodiment
Purpose, technical scheme and advantage for making the embodiment of the present invention are clearer, below in conjunction with the embodiment of the present invention
Middle accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described it is clear that described embodiment only
It is a part of embodiment of the present invention, rather than whole embodiments.Therefore, the enforcement to the present invention providing in the accompanying drawings below
The detailed description of example is not intended to limit the scope of claimed invention, but is merely representative of the selected enforcement of the present invention
Example.Based on embodiments of the invention, it is all that those skilled in the art are obtained on the premise of not making creative work
Other embodiment, broadly falls into the scope of protection of the invention.
Analysis of combining related cases is one of cracking of cases process important step, to serial case, multiple case, occupation
Sexual crime carries out series connection and merges, and can reduce and repeat to investigate, and saves human and material resources, the financial resource of investigation.The line in case
Rope, evidence are superimposed, a complete clue chain of connecting out, merge tissue investigation, it is possible to achieve detection together case and
Crack all by string and all cases.Mode of operation of investigating by combining related cases is detectd by combine related cases case source, confirmation case, development of discovery
Look into three stage compositions.For same subject of crime, during it is committed a crime continuously, its crime method, means meeting table
Reveal relative stability.Crime method, the relative stability of means, can reflect the internal relation of each case that it is located,
Represent its criminal activity rule and feature.Find, recognize this rule process be exactly investigate department series connection and case mistake
A series of cases are done the process of establishing identity by journey, are to find combine related cases source and the basis determining simultaneously case case.
At present, existing analytical technology of combining related cases, is by artificially specifying case feature, progressively collecting in multiple data bases
Similar case information, manually integrates to similar case information, then to substantial amounts of similar case, carries out case information successively
Analysis to judge with trajectory analysis.The work that is, existing public security is combined related cases relies primarily on repeated manual analyses, efficiency
Lowly.This analytical model workload of combining related cases is big, manual analyses high cost, time-consuming long, efficiency low it is difficult to when reply is information-based
The demand for development in generation.While additionally, this tradition is combined related cases, analytical model has influence on the speed solved a case and efficiency, with analysis
The workload of personnel increases, and the degree of fatigue of analysis personnel increases, and the accuracy of analysis also can be affected, and frequently can lead to point
Analysis not comprehensively, combine related cases not in time, part of should combining related cases fail string and or string going here and there simultaneously not in time, thus affecting the opportunity of solving a case adversely,
Cause serial case, criminal case to recur, lead to serious consequence.
Based on this, a kind of intelligent combine related cases analysis method and system provided in an embodiment of the present invention, existing string can be solved
And in case analytical model manual analyses lead to waste time and energy, the problem of inefficiency, it is possible to increase combine related cases analysis efficiency and
Accuracy rate.
For ease of understanding to the present embodiment, analysis method of first one kind disclosed in the embodiment of the present invention being combined related cases
Describe in detail,
Fig. 1 is that a kind of intelligence provided in an embodiment of the present invention is combined related cases the flow chart of analysis method.
With reference to Fig. 1, the method comprises the steps:
Step s102, based on police criminal detection Database case characteristic warehouse.
Specifically, set up the case characteristic warehouse based on big data, the data of police criminal detection data base is passed through number
It is drawn in case characteristic warehouse according to REPOSITORY TECHNOLOGY (extract-transform-load, abbreviation etl), prop up as data
Hold.Data warehouse (data warehouse) is a subject-oriented (subject oriented), integrated
(integrated) data acquisition system of, metastable (non-volatile), reflecting history change (time variant), uses
In support management decision-making (decision making support).
Etl is used for describing by data from source terminal through extracting (extract), conversion (transform), loading (load)
Process to destination.Etl, as the important link building data warehouse, extracts required data from data source.
Specifically, by distribution, the data in heterogeneous data source, such as relation data, flat data file etc. is drawn into interim intermediate layer
After be carried out, change, integrated, finally according to the data warehouse model pre-defining, load data into data warehouse or
In Data Mart, become the basis of on-line analytical processing, data mining.Etl is mainly used in building data warehouse and business originally
Smart items, are also increasingly being applied to migration, exchange and the synchronization of general information system data now.
Wherein, data pick-up is to extract, from source data origin system, the data that target data origin system needs.Carried by interface
Take source data, such as jdbc (java data base connectivity, java data base connects), specific database connection
With plane file extractor, and come extraction and its extracting mode of determination data with reference to source data;
The data obtaining from source data source according to business demand, is converted into the shape of target data source requirement by data conversion
Formula, and wrong, inconsistent data is carried out and processes;That is, the data that developer will extract, need according to business
Target data structure to be converted to, and realize collecting;
Data loads and for the data after conversion to be loaded into purpose data source.Load data that is converted and collecting to number of targets
According in warehouse, achievable sql (structured query language, SQL) or batch load.Etl is normal
Used in data warehouse, but its object is not limited to data warehouse.
In above 3 links of etl, how data pick-up, directly facing the data source of various dispersions, isomery, ensures steady
Determine efficiently to extract correct data from these data sources, be the key issue needing in etl design and implementation process to consider
One of.Data pick-up be according to actual scene Lai, generally, be first full dose extract, rear extended meeting is according to the actual requirements
To do increment extraction.When integrated end carries out the initialization of data, generally require and load by the total data of data source,
At this moment need to carry out full dose extraction.Full dose extracts similar to Data Migration or data duplication, and it is by the table in data source or view
Data all extract from data base, be converted into etl instrument (the oracle warehouse as oracle of oneself
The integration services of builder, sql server) form that can identify, then carry out follow-up loading operation.
Full dose is extracted and can be completed in the way of using data duplication, importing or backup, and realization mechanism is fairly simple.Full dose extracts and completes
Afterwards, the data having increased newly in table since follow-up extraction operation only need to be drawn from last time extraction or changing, here it is increment extraction.Mesh
The method of the conventional capture delta data in extracting of front incremental data has: trigger, timestamp, the contrast of full table, daily record contrast etc..
Which kind of increment extraction mechanism selected in etl implementation process actually, decision-making will be carried out according to actual data source systems environment, need
The type considering source system data storehouse, the data volume (determine severe) to performance requirement extracting, to source business system
The system control ability of database and realize the various factors such as difficulty, or even combine various different increment mechanism with for ring
The different data source systems in border carry out etl enforcement.
It is further that the data that in the present embodiment, etl extracts from police criminal detection data base includes case classification and (kills
People, robbery, theft etc.), case state (put on record, part of having solved a case, unsolved case etc.), case feature (live vestige, article etc.),
Personnel characteristics' (figure and features feature etc.), form case characteristic warehouse.Extract mode using using full dose first, when there being new case
Part produces, and needs to carry out increment extraction, and increment extraction mode is to be indexed to judge required increment with the time or with certain feature
Data, that is, based on timestamp, cooperation other modes carry out increment extraction.
Step s104, generates, by machine learning, analysis model of combining related cases based on case characteristic warehouse.
Specifically, based on case characteristic warehouse, by machine learning, the feature of the broken part of combining related cases of analysis, this spy
Levy including: crime means, crime feature, crime time, select the features such as place, generate combining related cases based on different Criminal characteristic
Analysis model is it is possible to actually used middle continuous renewal, perfect in user.That is, by case characteristic storehouse
Part information of solving a case in storehouse carries out machine learning, and formation is accurate, efficient, can automatically learn perfect analysis model of combining related cases.
Machine learning (machine learning, ml) is a multi-field cross discipline, be related to theory of probability, statistics,
The multi-door subjects such as Approximation Theory, convextiry analysis, algorithm complex theory.Specialize in the study how computer is simulated or realized the mankind
Behavior, to obtain new knowledge or skills, reorganizes existing knowledge structure and is allowed to constantly improve the performance of itself.It is people
The core of work intelligence, is to make computer have intelligent fundamental way, and throughout the every field of artificial intelligence, it is main for its application
Using conclusion, synthesis rather than deduction.The purpose of machine learning is exactly to allow machine have study similar to the mankind, understanding, reason
The ability of solution things.
It is further that machine learning includes step as shown in Figure 2:
S202: select training experience.
Specifically, the selection of the training experience providing to Learning machine has important impact for the success or failure of system.One
As for, training experience directly or indirectly should be able to make certain feedback to the decision-making of system, and training experience should
It is capable of the sequence of controlled training sample to a great extent;Additionally, training experience should also be as far as possible to training sample
Make good estimation with the spatial probability distribution of test sample.
S204: selection target function.
Specifically, after given training sample and training experience, Machine Learning Problems are just reduced to one and find ideal
The problem of object function f (x).
S206: the representation of selection target function.
Specifically, it is true that being to be stranded very much by preferable object function f (x) is obtained to the study of sample and training
Difficult, generally we are intended to obtain an approximate object function carry out the preferable object function of close approximation as far as possible.This
In close approximation, can be using the method such as quadratic polynomial function, neutral net realizing.
S208: select Function Approximation Algorithm.
Specifically, in order to obtain approximate objective function, we are according to the initial approximation object function selecting to training sample
Input estimated, thus obtain training sample estimation output, afterwards, using estimate output and reality output between mistake
Difference, to be fed back, is generally by weighed value adjusting.Then, the input to system reevaluates, and obtains new output and estimates
Evaluation, recalculates the error between estimated value and actual value, again system is fed back, and adjusts weight, is repeated in holding
Row is less than the threshold value setting or frequency of training more than the number of times setting until the total error of all training samples.
Step s106, using combining related cases, analysis model is analyzed to unsolved case, generates class case list.
Specifically, using analysis model of combining related cases, the unsolved case in case characteristic warehouse is analyzed, generates
Class case list.There is in Criminal characteristic the unsolved case of identical point, similitude, will act as class case and occur in such case list
In.It should be noted that analysis model of combining related cases can automatically analyze unsolved case feature, carry out classification analysis, generate class case
List is it is also possible to pass through the retrieval of outside (personnel tracked down by investigator, intelligence analysis personnel, case group leader etc. one line) (as case
Part feature) operation, generate class case list.
It is further, using analysis model analysis unsolved case of combining related cases, and when providing class case list, according to outside
The practical operation situation of (investigator, intelligence analysis personnel, case group leader etc. one line track down personnel), to the string of class case list simultaneously
Analysis, analysis is correct, is combined related cases analysis model by machine learning record;Deviation in analysis, by machine learning, constantly changes
Enter, perfect, continue to optimize the weight ratio of each dimension in analysis model of combining related cases, make this model of combining related cases further to reality
Application.
The intelligence of the embodiment of the present invention is combined related cases analysis method, by setting up the case characteristic storehouse based on big data
Storehouse, the data pick-up by etl for the data of police criminal detection data base forms in case characteristic warehouse, props up as data
Hold.Machine learning is passed through based on case characteristic warehouse, the feature of the broken part of combining related cases of analysis, generate special based on different crimes
The analysis model of combining related cases levied, is finally analyzed to unsolved case using analysis model of combining related cases, and generates class case list.Making
Pattern characteristics have the unsolved case of identical point, similitude, will act as class case and occur in such case list.This analysis of combining related cases
Model can pass through in actual use automatic study constantly update, perfect, further to real world applications, help public security
Criminal detective fast and effectively recommends possible combining related cases, and solves and needs when combining related cases analysis in prior art to investigate people
Member's manual analyses from police informatization search lead to waste time and energy, the problem of inefficiency, it is possible to increase combine related cases point
Analysis efficiency and accuracy rate.
Said method, in order to further improve the accuracy rate combined related cases, further comprising the steps of:
Step s108, sets up high-risk group data base based on public security previous conviction demographic data storehouse.
Specifically, by public security previous conviction personal information is carried out with cluster analyses, form high-risk group data base, can be to string
And case recommend may crime high-risk group.It is further that public security previous conviction demographic data storehouse includes plundering demographic data
Storehouse, drug addict data base, theft personnel's database.
Cluster analyses are the reasons according to " things of a kind come together, people of a mind fall into the same group ", a kind of multi-variate statistical analyses that sample or index are classified
Method, they discuss to as if substantial amounts of sample it is desirable to can reasonably by respective characteristic reasonably to be classified, do not have
Any pattern can be for reference or follow to carry out in the case of not having priori.Cluster is to sort data into not
Same class or such a process of cluster, so the object in same cluster has very big similarity, and right between different cluster
As there being very big diversity.The target of cluster analyses is exactly to collect data on the basis of similar to classify.Cluster comes from a lot
Field, including mathematics, computer science, statistics, biology and economics.In different applications, a lot of clustering techniques
It is developed, these technical methods are used as describing data, weigh the similarity between different data sources, and data source
It is categorized in different clusters.The computational methods that cluster analyses are commonly used mainly include several as follows: disintegrating method (partitioning
Methods), stratification (hierarchical methods), the method (density-based methods) based on density,
Method (grid-based methods) based on grid, the method (model-based methods) based on model.
In the present embodiment, cluster analyses, using based on algorithm model, select a class therein according to actual scene.
It should be noted that public security previous conviction demographic data storehouse is contained in case characteristic warehouse.
Step s110, sets up personnel activity's track database based on trace information.
Specifically, the personnel activity's track database based on big data, the track of the personnel that public security can be grasped are built
Information passes through etl data pick-up to personnel activity's track database.Wherein, the personnel that public security can be grasped include previous conviction personnel,
Personnel concerning the case, suspicion personnel.
It is further that trace information includes fence track, online track, lodging track, consumption track, public friendship
Pass-out row track.It should be noted that trace information is not limited to above-mentioned several track, also include ticket track, track of vehicle, ferrum
Rail mark, public transport track, subway track, civil aviaton track, bank track etc..Wherein, fence refers to mobile phone electronic fence,
It is also called wireless data acquisition terminal, belong to public safety bayonet type equipment, public security system special product.For public security and safety
Security department in the urgent need to and the electronic tracking of new generation of Development and Production controls equipment, using advanced mobile radio network and
Electronic information technology can carry out accurate management and control to specific region or the personnel specifying.Fence is by base station and daemon software
Two large divisions forms.Base station equipment has two kinds, and one kind is stationary base station, for being fixedly mounted on some outdoor environment fields for a long time
The mobile phone that institute's (public place, tourist attractions, critical facility, traffic intersection) comes in and goes out to specific region is monitored, another one
It is transportable type base station equipment, the interior such as hotel, Internet bar, ktv, bath center site can be deployed in, and buses, bus
The mobile spaces such as railway carriage or compartment, convenient, flexible.Daemon software is arranged on overall control center, according to the information of base station collection, by data analysiss
Data digging technology, can provide Monitoring Data in conjunction with relevant information such as identity card, license plate numbers for staff.
Fence can search the quantity of mobile phone by wireless signal, by base station can recognize that each mobile phone distance and
Track.Online track, the card registration of lodging track identity-based obtain, and consumption track is based on bank card information and obtains, public transport
Trip track includes, based on video monitoring, obtaining target initial information by Algorithm of Vehicle Detection by Video Image, calculating using average drifting
Method realizes the video tracking of target with reference to Kalman filtering algorithm, then its track is carried out with statistical analysiss, obtains the actual row of target
Sail the track characteristic in direction.
It should be noted that s108, s110 only for convenience of description, does not represent sequencing.
Step s112, transfers high-risk group's data from high-risk group data base, adjusts from personnel activity's track database
Take track data, using high-risk group's data and track data, class case list is carried out with data analysiss, generate list of combining related cases.
Specifically, analysis model of combining related cases accesses high-risk group data base, transfers high-risk people from high-risk group data base
Group's data;Analysis model of combining related cases accesses personnel activity's track database, transfers track number from personnel activity's track database
According to.
It should be noted that above transfer high-risk group's data and transfer not specific successively suitable between track data
Sequence, the two can be carried out with random order it is also possible to carry out simultaneously.
Using high-risk group's data and track data, class case list is carried out with data analysiss, generate list of combining related cases, enter one
Step improves the accuracy rate combined related cases.
The embodiment of the present invention additionally provides a kind of intelligent analysis system of combining related cases, and Fig. 3 is provided by the embodiment of the present invention
A kind of intelligence is combined related cases the structural framing figure of analysis system.
As shown in figure 3, this intelligence is combined related cases, analysis system includes data integration module 301, machine learning module 302 and class
Case generation module 303, data integration module 301 is connected with machine learning module 302, and machine learning module 302 is generated with class case
Module 303 is connected.
Wherein, data integration module 301 is used for based on police criminal detection Database case characteristic warehouse, engineering
Practise module 302 to be used for generating analysis model of combining related cases, class case generation module based on case characteristic warehouse by machine learning
303 are used for using analysis model of combining related cases, unsolved case being analyzed, and generate class case list.
This intelligence combine related cases analysis system specific work process as follows:
Data integration module 301 sets up the case characteristic warehouse based on big data, by the number of police criminal detection data base
It is drawn in case characteristic warehouse according to by etl, support as data.
Machine learning module 302, by carrying out machine learning, shape to the part information of solving a case in case characteristic warehouse
Become accurate, efficient, can automatically learn perfect analysis model of combining related cases.That is, machine learning module 302 is based on case spy
Levy data warehouse and pass through machine learning, the feature of the broken part of combining related cases of analysis, this feature includes: crime means, crime feature, work
The case time, select the feature such as place, generate analysis model of combining related cases based on different Criminal characteristic it is possible to reality in user
In use constantly update, perfect.Machine learning module 302 carries out the step of machine learning as shown in Fig. 2 will not be described here.
Class case generation module 303 is carried out to unsolved case using the analysis model of combining related cases that machine learning module 302 generates
Analysis, generates class case list.There is in Criminal characteristic the unsolved case of identical point, similitude, will act as class case and occur in this
In class case list.It should be noted that class case generation module 303 can automatically analyze unsolved case spy using model of combining related cases
Levy, sorted out, generate class case list it is also possible to receive an outside (line such as investigator, intelligence analysis personnel, case group leader
Detection personnel) retrieval (as case feature) signal, generate class case list.It is further that class case generation module 303 is using
When analysis model of combining related cases is analyzed unsolved case and provided class case list, according to outside (investigator, intelligence analysis personnel, case
Part group leader etc. one line track down personnel) practical operation signal, analysis result is fed back to machine learning module 302, for example, receives
Peripheral operation signal going here and there and analyzing correct to class case list, positive and negative for this result machine learning module 302 that is fed to is remembered by it
Record and update analysis model of combining related cases;Deviation in analysis, by this bias contribution negative feedback to machine learning module 302, continues
Continuous machine learning optimization is combined related cases the weight ratio of each dimension in analysis model, make this combine related cases model further to actual should
With.
Embodiments provide a kind of intelligent analysis system of combining related cases, including data integration module, machine learning mould
Block and class case generation module;Case characteristic warehouse based on police criminal detection data base, machine are set up by data integration module
Device study module is generated and is combined related cases analysis model by the solved a case feature of part of machine learning based on case characteristic warehouse, class
Case generation module is analyzed to unsolved case sorting out using analysis model of combining related cases, and generates class case list.In Criminal characteristic tool
There is the unsolved case of identical point, similitude, will act as class case and occur in such case list.This system addresses big data is analyzed
Technology is analyzed to public security industry, and depth excavates public business logic, and police criminal detection data base is imported by etl, is formed
Complete case characteristic warehouse, sets up a set of analysis model of combining related cases based on public security industry, thus solving information island,
Deep layer contact between depth mining data, provides the user useful decision information.Solve analysis of combining related cases in prior art
When need investigator from police informatization manual analyses search lead to waste time and energy, the problem of inefficiency, can
Improve combine related cases analysis efficiency and accuracy rate.
It is further that this intelligence analysis system of combining related cases also includes trajectory analysis module 304 and high-risk group and generates mould
Block 305.
Trajectory analysis module 304 is used for setting up personnel activity's track database based on trace information, and high-risk group generates mould
Block 305 is used for setting up high-risk group data base based on public security previous conviction demographic data storehouse.
Particularly, trajectory analysis module 304 extracts trace information by etl and sets up personnel activity's track database, rail
Mark information includes fence track, online track, lodging track, consumption track, transit trip track.High-risk group gives birth to
Become module 305 to be based on public security previous conviction demographic data storehouse, cluster analyses carried out to the delinquent personnel of solve a case part and correlation,
Depth excavates the relation of case feature and criminal feature, constructs high-risk group data base.
It should be noted that trajectory analysis module 304 and high-risk group's generation module 305 all with data integration module 301
It is connected, data integration module 301 is additionally operable to carry out the personal information in scattered trace information and previous conviction demographic data storehouse
Extract, integrate, trajectory analysis module 304 sets up personnel activity's track database based on the trace information after integrated, high
Danger crowd's generation module 305 is used for setting up high-risk group based on the integrated related personnel's information in public security previous conviction demographic data storehouse
Data base.
It is further that this intelligence analysis system of combining related cases also includes data access module 306.Track module 304, high-risk
Crowd's generation module 305 is all connected with data access module 306.The mode of data access includes two kinds, and one kind is Interworking Data
The mode in storehouse, another kind of by way of importing, such as excel form.
Preferably, data access module 306 includes data access port, accesses high-risk people by way of data is docked
Group database or personnel activity's track database.Data access module 306 is used for from personnel activity's track database, high-risk group
Personnel's track data, high-risk group's data is transferred respectively in data base.
Particularly, data access module 306 is from the high-risk group data base being set up by high-risk group's generation module 305
Transfer high-risk group's data, from the personnel activity's track database set up by trajectory analysis module 304, transfer track data.
It is further that this intelligence analysis system of combining related cases also includes analysis module 307.Data access module 306 with point
Analysis module 307 is connected, and analysis module 307 is connected with class case generation module 303.
Analysis module 307 is used for carrying out data analysiss using high-risk group's data and track data to class case list, generates
Combine related cases list.
Specifically, the high-risk people that analysis module 307 is transferred from high-risk group data base using data access module 306
Group's data and the track data transferred from personnel activity's track database, the class case row that class case generation module 303 is generated
Table carries out data analysiss and (generates, to class case list, the high-risk group that may commit a crime, and enter administrative staff, crime ground with reference to track data
The locus of points is analyzed), for public security, analysis personnel provide high-quality recommendation information of combining related cases, and generate list of combining related cases, subtract further
Few information in public security organs analyzes the workload of personnel.
The above, the only specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, and any
Those familiar with the art the invention discloses technical scope in, change or replacement can be readily occurred in, all should contain
Cover within protection scope of the present invention.Therefore, protection scope of the present invention should described be defined by scope of the claims.
Claims (10)
1. a kind of intelligence combines related cases analysis method it is characterised in that comprising the following steps:
Based on police criminal detection Database case characteristic warehouse;
Analysis model of combining related cases is generated by machine learning based on described case characteristic warehouse;
Using described analysis model of combining related cases, unsolved case is analyzed, generates class case list.
2. intelligence according to claim 1 combines related cases analysis method it is characterised in that further comprising the steps of:
High-risk group data base is set up based on public security previous conviction demographic data storehouse.
3. intelligence according to claim 2 combines related cases analysis method it is characterised in that further comprising the steps of:
Personnel activity's track database is set up based on trace information.
4. intelligence according to claim 3 combines related cases analysis method it is characterised in that further comprising the steps of:
Transfer high-risk group's data from described high-risk group data base;
Transfer track data from described personnel activity's track database;
Using described high-risk group's data and described track data, class case list is carried out with data analysiss, generate list of combining related cases.
5. intelligence according to claim 3 combines related cases analysis method it is characterised in that described trace information includes electronics encloses
Hurdle track, online track, lodging track, consumption track, transit trip track.
6. intelligence according to claim 1 combine related cases analysis method it is characterised in that described machine learning specifically include with
Lower step:
Select training experience;
Selection target function;
The representation of selection target function;
Select Function Approximation Algorithm.
7. a kind of intelligence combines related cases analysis system it is characterised in that including:
Data integration module, for based on police criminal detection Database case characteristic warehouse;
Machine learning module, for generating, by machine learning, analysis model of combining related cases based on described case characteristic warehouse;
Class case generation module, for being analyzed to unsolved case using described analysis model of combining related cases, generates class case list.
8. intelligence according to claim 7 combines related cases analysis system it is characterised in that also including:
Trajectory analysis module, for setting up personnel activity's track database based on trace information;
High-risk group's generation module, for setting up high-risk group data base based on public security previous conviction demographic data storehouse.
9. intelligence according to claim 8 combines related cases analysis system it is characterised in that also including:
Data access module, for transferring people respectively from described personnel activity's track database, described high-risk group data base
Member's track data, high-risk group's data;
Analysis module, for data analysiss being carried out to class case list using described high-risk group's data and described track data, raw
Become list of combining related cases.
10. intelligence according to claim 8 combines related cases analysis system it is characterised in that described trace information includes electronics
Fence track, online track, lodging track, consumption track, transit trip track.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610711740.5A CN106355537A (en) | 2016-08-23 | 2016-08-23 | Smart analysis method and system for interrelated cases |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610711740.5A CN106355537A (en) | 2016-08-23 | 2016-08-23 | Smart analysis method and system for interrelated cases |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106355537A true CN106355537A (en) | 2017-01-25 |
Family
ID=57844610
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610711740.5A Pending CN106355537A (en) | 2016-08-23 | 2016-08-23 | Smart analysis method and system for interrelated cases |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106355537A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108364438A (en) * | 2017-12-28 | 2018-08-03 | 武汉烽火众智数字技术有限责任公司 | Method for early warning based on front end perception data and device |
CN108874767A (en) * | 2018-05-04 | 2018-11-23 | 上海瀚所信息技术有限公司 | A kind of four dimension module intelligence Compare Systems and method for public security system |
CN109033351A (en) * | 2018-07-25 | 2018-12-18 | 北京神州泰岳软件股份有限公司 | The merging method and device of merit data |
CN109583669A (en) * | 2017-09-28 | 2019-04-05 | 北京国双科技有限公司 | Data capture method, device, storage medium and processor |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101187927A (en) * | 2007-12-17 | 2008-05-28 | 电子科技大学 | Criminal case joint investigation intelligent analysis method |
CN101814076A (en) * | 2009-07-29 | 2010-08-25 | 北京航天理想科技有限公司 | Visualizing information and information correlation analysis system and establishing method |
CN104572615A (en) * | 2014-12-19 | 2015-04-29 | 深圳中创华安科技有限公司 | Method and system for on-line case investigation processing |
CN104850905A (en) * | 2015-05-15 | 2015-08-19 | 上海集本信息技术服务有限公司 | Machine-learning-based legal risk quantitative evaluation system and method |
CN105069130A (en) * | 2015-08-14 | 2015-11-18 | 安徽新华博信息技术股份有限公司 | Suspect object prediction method |
CN105260449A (en) * | 2015-10-10 | 2016-01-20 | 张福辉 | Case key string serial-parallel detection method |
-
2016
- 2016-08-23 CN CN201610711740.5A patent/CN106355537A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101187927A (en) * | 2007-12-17 | 2008-05-28 | 电子科技大学 | Criminal case joint investigation intelligent analysis method |
CN101814076A (en) * | 2009-07-29 | 2010-08-25 | 北京航天理想科技有限公司 | Visualizing information and information correlation analysis system and establishing method |
CN104572615A (en) * | 2014-12-19 | 2015-04-29 | 深圳中创华安科技有限公司 | Method and system for on-line case investigation processing |
CN104850905A (en) * | 2015-05-15 | 2015-08-19 | 上海集本信息技术服务有限公司 | Machine-learning-based legal risk quantitative evaluation system and method |
CN105069130A (en) * | 2015-08-14 | 2015-11-18 | 安徽新华博信息技术股份有限公司 | Suspect object prediction method |
CN105260449A (en) * | 2015-10-10 | 2016-01-20 | 张福辉 | Case key string serial-parallel detection method |
Non-Patent Citations (3)
Title |
---|
孟剑萍等: "大数据助推公安信息化建设", 《微型机与应用》 * |
陈巍等: "基于案例推理的刑侦决策支持系统应用研究", 《山东警察学院学报》 * |
黄学鹏等: "警务信息综合应用平台研究与设计", 《计算机安全》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109583669A (en) * | 2017-09-28 | 2019-04-05 | 北京国双科技有限公司 | Data capture method, device, storage medium and processor |
CN108364438A (en) * | 2017-12-28 | 2018-08-03 | 武汉烽火众智数字技术有限责任公司 | Method for early warning based on front end perception data and device |
CN108874767A (en) * | 2018-05-04 | 2018-11-23 | 上海瀚所信息技术有限公司 | A kind of four dimension module intelligence Compare Systems and method for public security system |
CN108874767B (en) * | 2018-05-04 | 2022-04-05 | 上海瀚所信息技术有限公司 | Four-dimensional model intelligent comparison system and method for public security system |
CN109033351A (en) * | 2018-07-25 | 2018-12-18 | 北京神州泰岳软件股份有限公司 | The merging method and device of merit data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | Notice of retraction: intelligent transportation system in Macao based on deep self-coding learning | |
Demir et al. | Predicting occurrence of liquefaction-induced lateral spreading using gradient boosting algorithms integrated with particle swarm optimization: PSO-XGBoost, PSO-LightGBM, and PSO-CatBoost | |
CN107038505B (en) | Ore finding model prediction method based on machine learning | |
CN108596386A (en) | A kind of prediction convict repeats the method and system of crime probability | |
Xue et al. | Spatial analysis with preference specification of latent decision makers for criminal event prediction | |
CN104182805A (en) | Dangerous tendency prediction method based on prisoner behavior characteristic ensemble learning model | |
CN113157800A (en) | Identification method for discovering dynamic target in air in real time | |
CN106355537A (en) | Smart analysis method and system for interrelated cases | |
CN112800115B (en) | Data processing method and data processing device | |
CN106649809A (en) | Internet of Things analysis method | |
Xia et al. | A parallel grid-search-based SVM optimization algorithm on Spark for passenger hotspot prediction | |
CN108492561A (en) | A kind of road net traffic state Spatial And Temporal Characteristics method based on matrix decomposition | |
US11836637B2 (en) | Construction method of human-object-space interaction model based on knowledge graph | |
Cerqueti et al. | Municipal waste management: A complex network approach with an application to Italy | |
CN106777279A (en) | A kind of time-space relationship analysis system | |
Guan et al. | A unified framework for predicting kpis of on-demand transport services | |
Lei et al. | Mining spatial co-location patterns by the fuzzy technology | |
CN114418360A (en) | Smart city operation sign big data analysis method and device | |
Kumar et al. | Data mining based crime investigation systems: Taxonomy and relevance | |
CN111353085A (en) | Cloud mining network public opinion analysis method based on feature model | |
Pareek et al. | A review report on knowledge discovery in databases and various techniques of data mining | |
Patel et al. | Intellectual and enhance digital solution for police station | |
Li et al. | MGC-GAN: Multi-graph convolutional generative adversarial networks for accurate citywide traffic flow prediction | |
Gosavi et al. | A survey on crime occurrence detection and prediction techniques | |
Renushe et al. | Data mining practices for effective investigation of crime |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170125 |