CN107818175A - A kind of law class case problem intelligently prejudges system and method - Google Patents
A kind of law class case problem intelligently prejudges system and method Download PDFInfo
- Publication number
- CN107818175A CN107818175A CN201711142734.3A CN201711142734A CN107818175A CN 107818175 A CN107818175 A CN 107818175A CN 201711142734 A CN201711142734 A CN 201711142734A CN 107818175 A CN107818175 A CN 107818175A
- Authority
- CN
- China
- Prior art keywords
- data
- document
- database
- model
- feature
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/18—Legal services; Handling legal documents
Abstract
The present invention discloses a kind of law class case problem intelligence pre-judging method, comprises the following steps:Step A1, judgement document's database and property data base are set;Step S1, characteristic processing is carried out to first, second property data base, and the data after processing are stored in former feature database, obtain the precise information for subsequently modeling;Step S2, fisrt feature database data is imported as training set, builds model;Step S3, second feature storehouse data are imported as test set, above-mentioned gained model accuracy rate are tested, it is determined that final model;Step S4:User's legal issue is received, model is imported and is prejudged;Step S5:Anticipation result is sent to user.Such a method can be based on artificial intelligence technology destructing magnanimity document, and more direct, convenient, low price legal services are provided for vast legal requirements person, and providing law class case problem for numerous Internet users intelligently prejudges.Invention additionally discloses a kind of law class case problem intelligently to prejudge system.
Description
Technical field
It is more particularly to a kind of to utilize the big number of machine learning the present invention relates to computer, internet, machine learning field is belonged to
The system and method intelligently prejudged according to law class case problem is carried out.
Background technology
Reached its maturity with the theory and technology of artificial intelligence field, it has been achieved in terms of voice, image recognition
Mirable achievement, and legal field is for ground that artificial intelligence technology is still outside one piece of fresh generation set foot in less.
Under the background of internet rapid development, common people when running into dispute problem due to lack corresponding laws and regulations,
General knowledge of laws, similar cases, most suitable reply can not be made the very first time, if conventionally seeking help from lawyer, then
Expensive legal advice expense allows people to hang back.
Accumulative magnanimity document is a urgently bright-colored treasure-house leaved for development to legal field over the years, wherein what is contained is huge
Value attracts us and explored, and law is the science of a classification, in the legal case of same type, hides substantial amounts of
Homogeney condition, how type case to be used by artificial intelligence technology, or there is no the difficulty broken through in an industry
Topic.
The content of the invention
The purpose of the present invention, it is to provide a kind of law class case problem intelligently anticipation system and method, it can be based on artificial
Intellectual technology deconstructs magnanimity document, provides more direct, convenient, low price legal services for vast legal requirements person, is wide
Big Internet user provides law class case problem and intelligently prejudged.
In order to reach above-mentioned purpose, solution of the invention is:
A kind of law class case problem intelligently prejudges system, including:
One judgement document's database, judgement document's database include being used to store judgement document's case information and index
Judgement document storehouse;
One property data base, set a fisrt feature database, a second feature database in this feature database, for by
Ratio stores the document data after feature extraction, characteristic processing;
One feature extraction, processing unit, for carrying out feature extraction to the judgement document of judgement document's database, and by than
Example deposit fisrt feature database, second feature database:
One model creating unit, including:One model construction unit, it is connected with fisrt feature database, and it is special to import first
The data in database are levied, and the algorithm for combining corresponding data type is modeled, generation anticipation model;An and model measurement
Unit, it is connected with above-mentioned second feature database, the data imported in second feature database, to above-mentioned model construction unit
Gained model carries out accuracy rate test, to determine final model;And
One anticipation result transmitting element, including:One user port, for receiving user's legal issue;A kind of case prejudges mould
Block, the model finally determined based on above-mentioned model creating unit, corresponding judgement prediction is made to user's legal issue;And one is pre-
Sentence result sending port, anticipation result is sent to user.
Above-mentioned judgement document's database is also connected including one with judgement document storehouse, for data in active referee's document storehouse more
New document storehouse more new interface.
Above-mentioned fisrt feature database is used to store after feature extraction 80% document data, second feature database
For storing after feature extraction 20% document data.
Features described above database also includes a feature database more new interface being connected respectively with first, second property data base,
This feature storehouse more new interface is used for the renewal for performing data in first, second property data base in real time.
Features described above extraction, processing unit include:
One document characteristic extracting module, it is connected respectively with fisrt feature database, second feature database, for judge
The judgement document of document database carries out feature extraction, and is stored in above-mentioned first, second property data base in proportion;
One document feature processing block, it is connected respectively with fisrt feature database, second feature database, for above-mentioned
The data that property data base is stored in after the processing of document characteristic extracting module carry out characteristic processing, and will handle the data of completion again
It is stored back to first, second property data base.
A kind of law class case problem intelligence pre-judging method, comprises the following steps:
Step A1, judgement document's database and property data base are set;
Step A2, in judgement document's database advance typing judgement document information and using Reference Number as index and protect
Deposit;
Step A3, a fisrt feature database, a second feature database are set in the property data base;
Step A4, the class case keyword refined based on law team, is adopted piece by piece to judgement document in class case document database
Feature extraction is carried out with natural language processing technique;
Step A5, according to judgement document's total quantity, the document Jing Guo feature extraction is stored in first, second feature in proportion
Database is as training set and test set;
Step S1, characteristic processing is carried out to first, second property data base, and the data after processing are stored in former feature
Storehouse, obtain the precise information for subsequently modeling;
Step S2, fisrt feature database is imported as training set, it is suitable to be selected according to the data type of target signature
Algorithm fitting data builds model;
Step S3, second feature storehouse data are imported as test set, above-mentioned gained model accuracy rate is tested, according to
The height of accuracy rate determines it is to abandon model or reserving model, or the basis as Integrated Models, so that it is determined that final
Model;
Step S4:User's legal issue, including case type and case information are received, case information is disassembled into respective class
The characteristic information of case, and import in the final mask of above-mentioned preservation and prejudged;
Step S5:Anticipation result is sent to user.
In above-mentioned steps S1, the detailed process of characteristic processing includes:
Step S11, determines whether characteristic data type, each characteristic missing ratio and each feature have related pass
System, according to the different choice of three takes directly deletion, foundation intact part characteristic value is filled up or multivariate analysis skill
Art;
Step S12, logarithm value type data carry out surface analysis, the abnormity point beyond number range are detected, and determine
Retain or abandon;
Step S13, obtain final characteristic.
Above-mentioned steps S11 includes:It is determined that it is characterized in numeric type data or classifying type data, based on different data types
Take different treating methods;The shortage of data ratio of each characteristic is calculated, missing ratio directly deletes the spy more than 50%
Sign, the different choice based on missing pattern less than the ratio is to be filled up or selected multivariate analysis according to complete data
Method processing;The correlation of evaluation type data between any two is higher for missing ratio and the direct of correlated characteristic be present
Delete.
In above-mentioned steps S2, the data type for judging target signature is numeric type data or classifying type data, for number
Value type data decimation regression algorithm is modeled, and is modeled for classifying type data decimation sorting algorithm.
After such scheme, the present invention is on the basis of huge document storehouse and feature database documents, at natural language
Reason technology, machine learning techniques combination expertise structure class case intelligently prejudge system, the case information that each user is provided
Corresponding document characteristic information is disassembled into, by the way that its document characteristic information is imported into class case intelligently anticipation system, final acquisition is accurate
True decision in a case, the defects of existing attendant consultation service is costly is on the one hand overcome, on the other hand magnanimity document data are obtained
Utilized to effective.
Brief description of the drawings
Fig. 1 is the flow chart of the inventive method;
Fig. 2 is the Organization Chart of present system.
Embodiment
Below with reference to accompanying drawing, technical scheme and beneficial effect are described in detail.
As shown in Fig. 2 the present invention provides a kind of law class case problem intelligently anticipation system, including:
One judgement document's database 110, judgement document storehouse 111 is set in judgement document's database 110, judge for storing
Document case information and index;And the document storehouse more new interface 112 being connected with the judgement document storehouse 111, to perform in real time
The renewal of data in judgement document storehouse 111, include addition, deletion and the modification of judgement document storehouse Chinese letter breath;
One property data base 120, a fisrt feature database 121 is set in this feature database, pass through feature for storing
80% document data after extraction;One second feature database 122, for storing after feature extraction 20% document number
According to, and the feature database being connected with a described two property data bases more new interface 123, with perform in real time property data base 121,
In 122 in the renewal of data, including the property data base characteristic information addition, deletion and modification;
One feature extraction, processing unit 130, including:
One document characteristic extracting module 131, it is connected respectively with fisrt feature database 121, second feature database 122,
For carrying out feature extraction to the judgement document of judgement document's database, determined based on the specialized information that expert team provides corresponding
The keyword of class case, unified with nature language processing techniques carry out feature extraction one by one to judgement document, and are stored in proportion above-mentioned
First, second property data base;
One document feature processing block 132, it is connected respectively with fisrt feature database 121, second feature database 122,
For carrying out characteristic processing to the data that property data base is stored in after the processing of above-mentioned document characteristic extracting module 131, system is included
The methods of counting missing values processing or the outlier processing in method, is respectively used to fill up the data of missing, ensures the complete of data
Abnormity point in whole property and removal data, ensures the accuracy of data, the accuracy of data, which can significantly affect, subsequently establishes mould
The accuracy rate of type, therefore characteristic processing is important;And the data for handling completion are stored back to first, second property data base again;
One model creating unit 140, including:
One model construction unit 141, it is connected, is imported in fisrt feature database with above-mentioned fisrt feature database 121
Data are modeled, and it is classifying type data or numeric type data to first confirm that data type, and the difference based on data type is selected
Classification or regression algorithm are selected, above-mentioned data are fitted by algorithm and are modeled, generation anticipation model;
One model measurement unit 142, it is connected, is imported in second feature database with above-mentioned second feature database 122
Data carry out accuracy rate test to the above-mentioned gained model of model construction unit 141, and performance of the observing and nursing on test set is with certainly
It is fixed whether reserving model.
One anticipation result transmitting element 150, including:
One user port 151, for receiving user's legal issue;User can by internet to anticipation system input method
Rule problem, including legal issue classification and correlated characteristic information;
A kind of case prejudges module 152, and the class case finally determined based on above-mentioned model creating unit 140 prejudges model, to
Family legal issue makes corresponding judgement prediction.
One anticipation result sending port 153, anticipation result is sent to user.
Coordinate shown in Fig. 1, the present invention also provides a kind of law class case problem intelligence pre-judging method, comprises the following steps:
Step A1, judgement document's database and property data base are set;
Step A2, in judgement document's database advance typing judgement document information and using Reference Number as index and protect
Deposit;
Judgement document manual entry according to the actual requirements, and this article stack room can pass through connected document storehouse
More new interface is updated at any time, includes addition, deletion and the modification of document;
Step A3, a fisrt feature database, a second feature database are set in the property data base;
Step A4, the class case keyword refined based on law team, is adopted piece by piece to judgement document in class case document database
Feature extraction is carried out with natural language processing technique;
Step A5, according to judgement document's total quantity, the document Jing Guo feature extraction 80% is stored in fisrt feature database
Second feature database is stored in as test set as training set, remaining 20% document, and first, second property data base can
To be updated at any time by corresponding more new interface, include addition, deletion and the modification of corresponding feature;
As can be seen that above-mentioned steps A1~A5 is the primitive accumulation process of a data, once class case judgement document or phase
Feature input database is answered, before its needs is updated, it is no longer necessary to any human intervention, equivalent to disposable typing;
Step S1:Next, carrying out characteristic processing to first, second property data base, and the data after processing are stored in
Former feature database, obtain the precise information for subsequently modeling;
The detailed process of the characteristic processing includes:
Step S11, missing values processing;
Characteristic data type is determined, each characteristic lacks ratio, and whether each feature has dependency relation etc., according to three
The different choice of person takes directly deletion, filled up according to intact part characteristic value or the processing such as multivariate analytical techniques is done
Method;
Specifically, missing values processing may comprise steps of:
It is determined that being characterized in numeric type data or classifying type data, different processing is taken to do based on different data types
Method;
The shortage of data ratio of each characteristic is calculated, missing ratio directly deletes this feature more than 50%, less than the ratio
The different choice based on missing pattern of example is to be filled up according to complete data or select Multivariate to handle;
The correlation of evaluation type data between any two is higher for missing ratio and can selecting for correlated characteristic be present
Select direct deletion;
Step S12, outlier processing;
Logarithm value type data carry out surface analysis, and the abnormity point beyond number range is detected, and determine to retain still
Abandon
Step S13, obtain final characteristic;
Step S2, model construction:Fisrt feature database is imported as training set, determining the data type of target signature is
Numeric type data or classifying type data, suitable algorithm fitting data structure model is selected, specifically, for numeric type number
It is modeled according to regression algorithm is chosen, is modeled for classifying type data decimation sorting algorithm;
Step S3, model measurement:Second feature storehouse data are imported as test set, above-mentioned gained model accuracy rate is carried out
Test, determine it is to abandon model or reserving model according to the height of accuracy rate, or the basis as Integrated Models, so as to really
Fixed final model;
Step S4:User's legal issue, including case type and case information are received, case information is disassembled into respective class
The characteristic information of case, and import in the final mask of above-mentioned preservation and prejudged;
Step S5:Anticipation result is sent to user.
The technological thought of above example only to illustrate the invention, it is impossible to protection scope of the present invention is limited with this, it is every
According to technological thought proposed by the present invention, any change done on the basis of technical scheme, the scope of the present invention is each fallen within
Within.
Claims (9)
1. a kind of law class case problem intelligently prejudges system, it is characterised in that including:
One judgement document's database, judgement document's database include being used for the judge for storing judgement document's case information and index
Document storehouse;
One property data base, a fisrt feature database, a second feature database is set in this feature database, in proportion
Store the document data after feature extraction, characteristic processing;
One feature extraction, processing unit, for carrying out feature extraction to the judgement document of judgement document's database, and deposit in proportion
Enter fisrt feature database, second feature database:
One model creating unit, including:One model construction unit, it is connected with fisrt feature database, imports fisrt feature number
According to the data in storehouse, and the algorithm for combining corresponding data type is modeled, generation anticipation model;An and model measurement list
Member, it is connected with above-mentioned second feature database, the data imported in second feature database, to above-mentioned model construction unit institute
Obtain model and carry out accuracy rate test, to determine final model;And
One anticipation result transmitting element, including:One user port, for receiving user's legal issue;A kind of case prejudges module, base
In the model that above-mentioned model creating unit finally determines, corresponding judgement prediction is made to user's legal issue;And one anticipation knot
Fruit sending port, anticipation result is sent to user.
2. a kind of law class case problem as claimed in claim 1 intelligently prejudges system, it is characterised in that:Judgement document's number
Also include a document storehouse more new interface for being connected with judgement document storehouse, being updated for data in active referee's document storehouse according to storehouse.
3. a kind of law class case problem as claimed in claim 1 intelligently prejudges system, it is characterised in that:The fisrt feature number
It is used to store after feature extraction 80% document data according to storehouse, second feature database is used to store after feature extraction
20% document data.
4. a kind of law class case problem as claimed in claim 1 intelligently prejudges system, it is characterised in that:The property data base
Also include a second more new interface being connected respectively with first, second property data base, the second more new interface is used to hold in real time
The renewal of data in the property data base of row first, second.
5. a kind of law class case problem as claimed in claim 1 intelligently prejudges system, it is characterised in that:The feature extraction,
Processing unit includes:
One document characteristic extracting module, it is connected respectively with fisrt feature database, second feature database, for judgement document
The judgement document of database carries out feature extraction, and is stored in above-mentioned first, second property data base in proportion;
One document feature processing block, it is connected respectively with fisrt feature database, second feature database, for above-mentioned document
The data that property data base is stored in after characteristic extracting module processing carry out characteristic processing, and the data for handling completion are stored back to again
First, second property data base.
6. a kind of law class case problem intelligence pre-judging method, it is characterised in that comprise the following steps:
Step A1, judgement document's database and property data base are set;
Step A2, in judgement document's database advance typing judgement document information and using Reference Number as index and preserve;
Step A3, a fisrt feature database, a second feature database are set in the property data base;
Step A4, the class case keyword refined based on law team, judgement document in class case document database is used certainly piece by piece
Right language processing techniques carry out feature extraction;
Step A5, according to judgement document's total quantity, the document Jing Guo feature extraction is stored in first, second characteristic in proportion
Storehouse is as training set and test set;
Step S1, characteristic processing is carried out to first, second property data base, and the data after processing are stored in former feature database, obtained
The precise information subsequently modeled must be used for;
Step S2, fisrt feature database is imported as training set, suitable algorithm is selected according to the data type of target signature
Fitting data builds model;
Step S3, second feature storehouse data are imported as test set, above-mentioned gained model accuracy rate are tested, according to accurate
The height of rate determines it is to abandon model or reserving model, or the basis as Integrated Models, so that it is determined that final model;
Step S4:User's legal issue, including case type and case information are received, case information is disassembled into respective class case
Characteristic information, and import in the final mask of above-mentioned preservation and prejudged;
Step S5:Anticipation result is sent to user.
A kind of 7. law class case problem intelligence pre-judging method as claimed in claim 6, it is characterised in that:In the step S1,
The detailed process of characteristic processing includes:
Step S11, determines whether characteristic data type, each characteristic missing ratio and each feature have dependency relation, root
According to the different choice of three takes directly deletion, foundation intact part characteristic value is filled up or multivariate analytical techniques;
Step S12, logarithm value type data carry out surface analysis, the abnormity point beyond number range are detected, and determine to retain
Or abandon;
Step S13, obtain final characteristic.
A kind of 8. law class case problem intelligence pre-judging method as claimed in claim 7, it is characterised in that:The step S11 bags
Include:It is determined that being characterized in numeric type data or classifying type data, different treating methods is taken based on different data types;Meter
The shortage of data ratio of each characteristic is calculated, missing ratio directly deletes this feature more than 50%, less than the ratio based on scarce
The different choice of disabling mode is to be filled up according to complete data or select Multivariate to handle;Evaluation type number
According to correlation between any two, direct deletion that is higher for missing ratio and correlated characteristic being present.
A kind of 9. law class case problem intelligence pre-judging method as claimed in claim 6, it is characterised in that:In the step S2,
The data type for judging target signature is numeric type data or classifying type data, chooses regression algorithm for numeric type data and enters
Row modeling, is modeled for classifying type data decimation sorting algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711142734.3A CN107818175B (en) | 2017-11-17 | 2017-11-17 | Legal case problem analysis method and device based on referee document |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711142734.3A CN107818175B (en) | 2017-11-17 | 2017-11-17 | Legal case problem analysis method and device based on referee document |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107818175A true CN107818175A (en) | 2018-03-20 |
CN107818175B CN107818175B (en) | 2021-08-24 |
Family
ID=61610034
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711142734.3A Active CN107818175B (en) | 2017-11-17 | 2017-11-17 | Legal case problem analysis method and device based on referee document |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107818175B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110059079A (en) * | 2019-04-28 | 2019-07-26 | 北京深醒科技有限公司 | A kind of personnel based on big data modeling analysis break laws and commit crime prediction technique and system |
CN110297989A (en) * | 2019-06-25 | 2019-10-01 | 北京百度网讯科技有限公司 | Test method, device, equipment and the medium of abnormality detection |
CN110738032A (en) * | 2018-07-03 | 2020-01-31 | 北京国双科技有限公司 | method and device for generating judgment sections of official paperwork institute |
CN110858269A (en) * | 2018-08-09 | 2020-03-03 | 清华大学 | Criminal name prediction method and device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201804333U (en) * | 2010-06-18 | 2011-04-20 | 上海百事通信息技术有限公司 | Intelligent diagnostic system for legal issues |
US20140058960A1 (en) * | 2012-08-27 | 2014-02-27 | Datacert, Inc. | Predictive modeling: litigation decision analysis and optimization |
CN106599815A (en) * | 2016-12-06 | 2017-04-26 | 东南大学 | Mark distribution based head posture estimation method solving problem of class deletion |
CN106933787A (en) * | 2017-03-20 | 2017-07-07 | 上海智臻智能网络科技股份有限公司 | Adjudicate the computational methods of document similarity, search device and computer equipment |
CN107066468A (en) * | 2016-11-18 | 2017-08-18 | 北京市高级人民法院 | A kind of case search method based on genetic algorithm and nearest neighbor algorithm |
CN107194428A (en) * | 2017-05-26 | 2017-09-22 | 重庆师范大学 | A kind of integrated self-training method based on neighbour's density and semi-supervised KNN |
CN107291688A (en) * | 2017-05-22 | 2017-10-24 | 南京大学 | Judgement document's similarity analysis method based on topic model |
-
2017
- 2017-11-17 CN CN201711142734.3A patent/CN107818175B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201804333U (en) * | 2010-06-18 | 2011-04-20 | 上海百事通信息技术有限公司 | Intelligent diagnostic system for legal issues |
US20140058960A1 (en) * | 2012-08-27 | 2014-02-27 | Datacert, Inc. | Predictive modeling: litigation decision analysis and optimization |
CN107066468A (en) * | 2016-11-18 | 2017-08-18 | 北京市高级人民法院 | A kind of case search method based on genetic algorithm and nearest neighbor algorithm |
CN106599815A (en) * | 2016-12-06 | 2017-04-26 | 东南大学 | Mark distribution based head posture estimation method solving problem of class deletion |
CN106933787A (en) * | 2017-03-20 | 2017-07-07 | 上海智臻智能网络科技股份有限公司 | Adjudicate the computational methods of document similarity, search device and computer equipment |
CN107291688A (en) * | 2017-05-22 | 2017-10-24 | 南京大学 | Judgement document's similarity analysis method based on topic model |
CN107194428A (en) * | 2017-05-26 | 2017-09-22 | 重庆师范大学 | A kind of integrated self-training method based on neighbour's density and semi-supervised KNN |
Non-Patent Citations (2)
Title |
---|
COLLINS 等: ""An application of a multiple neural network learning system to emulation of mortgage underwriting judgements"", 《IEEE 1988 INTERNATIONAL CONFERENCE ON NEURAL NETWORKS》 * |
刘萍: ""基于大数据背景的经济法律案件智能预判系统结构分析"", 《经济与法》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110738032A (en) * | 2018-07-03 | 2020-01-31 | 北京国双科技有限公司 | method and device for generating judgment sections of official paperwork institute |
CN110738032B (en) * | 2018-07-03 | 2024-02-13 | 北京国双科技有限公司 | Method and device for generating judge paperwork thinking section |
CN110858269A (en) * | 2018-08-09 | 2020-03-03 | 清华大学 | Criminal name prediction method and device |
CN110858269B (en) * | 2018-08-09 | 2022-03-08 | 清华大学 | Fact description text prediction method and device |
CN110059079A (en) * | 2019-04-28 | 2019-07-26 | 北京深醒科技有限公司 | A kind of personnel based on big data modeling analysis break laws and commit crime prediction technique and system |
CN110297989A (en) * | 2019-06-25 | 2019-10-01 | 北京百度网讯科技有限公司 | Test method, device, equipment and the medium of abnormality detection |
CN110297989B (en) * | 2019-06-25 | 2021-07-23 | 北京百度网讯科技有限公司 | Test method, device, equipment and medium for anomaly detection |
Also Published As
Publication number | Publication date |
---|---|
CN107818175B (en) | 2021-08-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107818175A (en) | A kind of law class case problem intelligently prejudges system and method | |
CN109783632A (en) | Customer service information-pushing method, device, computer equipment and storage medium | |
CN107862022B (en) | Culture resource recommendation system | |
CN107851097A (en) | Data analysis system, data analysing method, data analysis program and storage medium | |
CN107342976A (en) | For the mobile solution platform and method of enterprise's Analysis on Industry Chain | |
CN110119877B (en) | Target employee selection method and device | |
CN111522733B (en) | Crowdsourcing tester recommending and crowdsourcing testing method and electronic device | |
CN107093091B (en) | Data processing method and device | |
CN106934010A (en) | Automatic duplicate checking method and device | |
CN109711424B (en) | Behavior rule acquisition method, device and equipment based on decision tree | |
KR20180072167A (en) | System for extracting similar patents and method thereof | |
CN106708940A (en) | Method and device used for processing pictures | |
CN109508420A (en) | A kind of cleaning method and device of knowledge mapping attribute | |
CN108846695A (en) | The prediction technique and device of terminal replacement cycle | |
CN108897732A (en) | Statement type recognition methods and device, storage medium and electronic device | |
CN108920677A (en) | Questionnaire method, investigating system and electronic equipment | |
CN107918825A (en) | A kind of method and apparatus that age of user section is judged based on application installation preference | |
CN109978020A (en) | A kind of social networks account vest identity identification method based on multidimensional characteristic | |
CN108509597A (en) | Work mark registers power evaluation method and system | |
CN105786910B (en) | Entry weighing computation method and device | |
Díaz et al. | A comparative approach between different computer vision tools, including commercial and open-source, for improving cultural image access and analysis | |
CN106909534A (en) | A kind of method and device for differentiating text-safe | |
CN109408396A (en) | Method for evaluating software quality, device, equipment and computer readable storage medium | |
CN107704763A (en) | Multi-source heterogeneous leak information De-weight method, stage division and device | |
CN107430590A (en) | Data compare |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |