CN110222866A - In conjunction with the intelligent civil case forecasting system and method for colloquial style description and question and answer - Google Patents
In conjunction with the intelligent civil case forecasting system and method for colloquial style description and question and answer Download PDFInfo
- Publication number
- CN110222866A CN110222866A CN201910348700.2A CN201910348700A CN110222866A CN 110222866 A CN110222866 A CN 110222866A CN 201910348700 A CN201910348700 A CN 201910348700A CN 110222866 A CN110222866 A CN 110222866A
- Authority
- CN
- China
- Prior art keywords
- feature
- user
- colloquial style
- question
- merit
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000013589 supplement Substances 0.000 claims abstract description 25
- 238000001514 detection method Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000000295 complement effect Effects 0.000 claims description 7
- 238000013136 deep learning model Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000009223 counseling Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- GOLXNESZZPUPJE-UHFFFAOYSA-N spiromesifen Chemical compound CC1=CC(C)=CC(C)=C1C(C(O1)=O)=C(OC(=O)CC(C)(C)C)C11CCCC1 GOLXNESZZPUPJE-UHFFFAOYSA-N 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q50/18—Legal services
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Technology Law (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of intelligent civil case prediction techniques of combination colloquial style description and question and answer, comprising the following steps: S1. receives the colloquial style merit description of user's input;S2. describe to determine that the consulting of user is intended to according to the colloquial style merit;S3. whether complete detection feature is intended to according to the consulting, if so, thening follow the steps S4, otherwise user is prompted to supplement corresponding feature;S4. prediction model is called to export corresponding consulting result to user according to complete feature.It is changed into based on being based on describing with colloquial style by the form of pure question and answer, a small amount of problem is supplemented and the form combined reduces the quantity that user needs passively to answer a question, and promotes user experience and completion rate.
Description
Technical field
The invention belongs to legal services technical field more particularly to a kind of description of combination colloquial style are civil with the intelligence of question and answer
Case forecasting system and method.
Background technique
As country's promotion society governed by law and citizen's pay attention to day by day safeguard the equity of oneself, need of the society to legal services
Ask also more and more.It is domestic at present to be serviced only with respect to the open of civil case online consultation, with the professional consultation in civil field
For, including following steps, as shown in Figure 1:
1) user's select permeability type.You are by one problem types of selection herein, to limit the range of consulting.It is typical
The problem of type be some claims type under some dispute type.
2) system pops up a series of problems, and these problems are combed by law business expert, the answer corresponding one of each problem
A template or legal provision.
3) user answers a series of problems, and user, which needs to answer, some selectes corresponding under problem types ask in 1)
Topic, these problems are arranged by law business expert, and the problem quantity and content under particular problem type are fixed and invariable.
4) the corresponding template of matching question and answer according to preset template and combines answer of the user to problem, provides consulting report
It accuses, consultation report is summarizing for multiple portions, each customer problem can correspond to a paragraph of a consultation report.
Above system has the disadvantage in that
1. user needs to pre-select specific problem types, but due to by culture and education degree and to legal field profession
The influence of vocabulary degree of understanding will lead to user and select difficult or selection error.Selection is difficult, refers to user for that should select
There is no a specific in-mind anticipations for which most suitable problem types, or too many due to selecting, and user experience is very poor;Choosing
Error is selected, refers to that user has selected a problem type unrelated with the problem of seeking advice from type oneself is needed after thinking, leads
Following user is caused to answer some problems unrelated with oneself problem types.
2. not supporting the colloquial description i.e. input of natural language, the input of user is based entirely on single choice or the side of multiselect
The question and answer mode of formula, system can only passively collect the problem of preparing in advance corresponding answer, but since legal issue range is wide
General, customer problem personalized reason, even if after user answers series of questions, the problem of user needs most consulting or user
Key point deeply concerned still may not be collected into.The unicity for collecting information phase mode, causes information input unilateral, into
And influence the quality of case consulting.And under real scene, commission of jurists listens attentively to the description of user first, then has what is be directed to mention
It asks, so that it is determined that case and candidate legislative advice.
3. the generation for seeking advice from result can only cannot rely on mass data and based on machine learning and depth by artificial combing
Spend the learning ability of study.
The generation of consultation report needs the manpower work of a large amount of legal professionals based entirely on the rule manually combed
Make;Consultation report can be provided, and without the powerful learning ability using machine, to also not embody intelligent capability.
4. needing to inquire the very more problem of user, the ratio for causing user to complete entire consultation process is very low.Problem is not
Can according to the user's choice, dynamic display and the maximally related problem of user's merit, the user of each selection same problem type
In face of the problem of be the same.
Although above-mentioned online open counseling services have many advantages, such as convenient, low price, far from due to previous reasons
And the counseling services of true lawyer, and true lawyer's limited amount, and consultant expenses is high, and ordinary user is caused still cannot
Obtaining has specific aim and low-cost counseling services.
Summary of the invention
Regarding the issue above, the present invention provides a kind of intelligent civil case of combination colloquial style description and question and answer
Prediction technique;
It is another object of the present invention in view of the above-mentioned problems, to provide a kind of description of combination colloquial style civil with the intelligence of question and answer
Case forecasting system.
In order to achieve the above objectives, present invention employs following technical proposals:
The intelligent civil case prediction technique of combination colloquial style description and question and answer of the invention, comprising the following steps:
S1. the colloquial style merit description of user's input is received;
S2. describe to determine that the consulting of user is intended to according to the colloquial style merit;
S3. whether complete detection feature is intended to according to the consulting, if so, thening follow the steps S4, otherwise prompt is used
Family supplements corresponding feature;
S4. prediction model is called to export corresponding consulting result to user according to complete feature.
It is in step sl, described in the intelligent civil case prediction technique of above-mentioned combination colloquial style description and question and answer
Colloquial style merit description include literal type merit description and sound-type merit describe, and when for sound-type case
When feelings describe, first passes through the speech recognition system and convert speech into text.
It is in step s 2, described in the intelligent civil case prediction technique of above-mentioned combination colloquial style description and question and answer
Consulting is intended to include problem types and dispute main body, and described problem type includes dispute type and demand.
In the intelligent civil case prediction technique of above-mentioned combination colloquial style description and question and answer, in step s 2, pass through
Following manner determines that consulting is intended to:
S21. dispute type and demand are identified using intention assessment engine, and judges whether are the dispute type and demand
Meet confidence level requirement, if so, thening follow the steps S22, the candidate item for otherwise generating dispute type and demand type is selected for user
It selects;
S22. type of subject is identified using intention assessment engine and judge whether the type of subject meets confidence level requirement,
If so, thening follow the steps S3, otherwise, type of subject candidate item is generated according to the dispute type and demand and is selected for user.
In the intelligent civil case prediction technique of above-mentioned combination colloquial style description and question and answer, in step s3, detection
The whether complete method of feature includes:
S31. it is intended to determine all characteristic informations to be obtained according to the consulting;
S32. the colloquial style merit is handled using feature calculation engine to describe, and look in colloquial style merit description
To the feature for matching the characteristic information;
S33. judge with the presence or absence of non-matching characteristic in all characteristic informations, if so, testing result is characterized content not
Completely.
In the intelligent civil case prediction technique of above-mentioned combination colloquial style description and question and answer, in step s3, pass through
Following methods prompt user to supplement corresponding feature:
S34. all non-matching characteristics are ranked up according to feature importance;
S35. the ranking results are made into candidate problem list, and prompt user to supplement phase in a manner of puing question to user
The feature answered.
In the intelligent civil case prediction technique of above-mentioned combination colloquial style description and question and answer, step S4 is specifically included:
S41. the prediction model is won a lawsuit the probability value of probability according to the output of all features for demand;
S42. obtain merit and predict reason template, and according to all features and probability value output comprising prediction result and
Predict the consulting result of reason.
In the intelligent civil case prediction technique of above-mentioned combination colloquial style description and question and answer, the prediction model passes through
Acquisition is trained to deep learning model using judgement document's data by mark.
A kind of intelligent civil case forecasting system of combination colloquial style description and question and answer, including preprocessing module, prediction mould
Block and output module, wherein
Preprocessing module, whether comprising complete feature in the colloquial style merit description for detecting user's input,
If imperfect, user is prompted to supplement;
Prediction module, for calling prediction model and being won a lawsuit probability based on the output demand of complete feature;
Output module, for being based on merit prediction reason according to the prediction result of the prediction module and complete feature
Consulting result is exported by template.
In the intelligent civil case forecasting system of above-mentioned combination colloquial style description and question and answer, the preprocessing module
Including intention assessment engine, feature identification engine and problem complementary module, wherein
Intention assessment engine, the colloquial style merit for being inputted according to user describe to determine that the consulting of user is intended to;
Feature identifies engine, all characteristic informations to be obtained is determined for being intended to according to the consulting, in colloquial style case
The feature for matching the characteristic information is found in feelings description;
Problem complementary module, for non-matching characteristic to be made into candidate problem to prompt user to supplement in corresponding feature
Hold.
The intelligent civil case forecasting system and method for combination colloquial style description and question and answer of the invention, are based on deep learning
Technology turns to data-driven, You Chunwen from regular drive using the deep learning model by the training of magnanimity judgement document data
The form answered is changed into based on based on being described with colloquial style, and a small amount of problem supplement and the form that combines reduces user and needs passive time
The quantity of question and answer topic, promotes user experience and completion rate.
Detailed description of the invention
Fig. 1 is the work flow diagram of the legal advice system of the prior art;
Fig. 2 is the method flow diagram that civil case is predicted in the embodiment of the present invention one;
Fig. 3 is the workflow schematic diagram of intention assessment engine in the embodiment of the present invention one;
Fig. 4 is the method flow diagram that feature is detected in the embodiment of the present invention one;
Fig. 5 is the workflow schematic diagram of feature calculation engine in the embodiment of the present invention one;
Fig. 6 is consulting result generating process flow chart in the embodiment of the present invention one;
Fig. 7 is the system structure diagram of civil case forecasting system in the embodiment of the present invention two.
In figure, preprocessing module 1;Intention assessment engine 11;Feature identifies engine 12;Problem complementary module 13;Predict mould
Block 2;Output module 3.
Specific embodiment
The Supreme People's Court in 2013 sets up Chinese judgement document's net, and the disclosure of mass data is so that law data resource
Change;The knowledge mapping of legal field effectively organizes various isolated information with the addressable information of unification, makes
Information use and the reasoning for obtaining legal field are more intelligent;The development of machine learning especially depth learning technology, it is automatical and efficient
Slave big data in learnt so that machine has an opportunity to complete specific task in the form of certain intelligence, utilize law
Machine learning techniques will be applied to legal field by mass data to have broad application prospects.The present invention is based on depth
Habit technology turns to data-driven from regular drive, by pure using the deep learning model by the training of magnanimity judgement document data
The form of question and answer is changed into based on based on being described with colloquial style, and the form of a small amount of problem supplement, which simultaneously combines, to be reduced user and need to answer
The quantity of problem promotes user experience and completion rate.
It is the preferred embodiment of the present invention and in conjunction with attached drawing below, technical scheme of the present invention will be further described,
However, the present invention is not limited to these examples.
Embodiment one
As shown in Fig. 2, present embodiment discloses a kind of intelligent civil case prediction sides of combination colloquial style description and question and answer
Method, comprising the following steps:
S1. the colloquial style merit description of user's input is received;
S2. describe to determine that the consulting of user is intended to according to the colloquial style merit;
S3. whether complete detection feature is intended to according to the consulting, if so, thening follow the steps S4, otherwise prompt is used
Family supplements corresponding feature;
S4. prediction model is called to export corresponding consulting result to user according to complete feature.
Specifically, in step sl, colloquial style merit description includes the merit description and voice class of literal type
The merit of type describes, and when describing for the merit of sound-type, first pass through the speech recognition system digitize the speech into it is written
This.The description of great convenience user and selection merit letter in such a way that the colloquial style that text and voice input is in conjunction with question and answer
Breath, at the same it is very convenient in such a way that text and voice input natural language, it also reduces the not high people of education level and consults
Ask the degree of difficulty of legal issue.
It, can not be into the case where user's colloquial style describes more sufficient situation, that is, in the complete situation of feature
Row problem supplement, that is, do not have question and answer link, in the case where user's colloquial style describes insufficient situation, carries out problem supplement, further mentions
The validity of case prediction is risen.
Specifically, prediction model is trained deep learning model by using judgement document's data by mark and obtains
?.Those skilled in the art would know how to be trained to be wanted deep learning model using sample database data
Model complete specific task, detailed process is not repeated herein, simply introduced training process below: first
Then the judgement document's data for extracting magnanimity distinguish different cases by, demand type, labeling system are used to complete judgement structure
Automation mark, and mark personnel is made to adjudicate document by information extractions such as, demand types according to different cases in labeling system
In feature, and realize by successive ignition mode the data mark of high-accuracy, use the later judgement document's data of mark
Deep learning model is trained.The input of training data is the case description in judgement document, and output is in specific case
The probability of plaintiff's demand is supported by judgement under, demand type.
Specifically, as shown in figure 3, in step s 2, consulting is intended to include problem types and dispute main body, and problem types are logical
Often refer to the demand under dispute type and dispute type.And determine that consulting is intended in the following manner:
S21. dispute type and demand are identified using intention assessment engine, and judges whether are the dispute type and demand
Meet confidence level requirement, if so, thening follow the steps S22, the candidate item for otherwise generating dispute type and demand type is selected for user
It selects, user selects a specific dispute type and at least one demand from candidate item;
S22. type of subject is identified using intention assessment engine and judge whether the type of subject meets confidence level requirement,
If so, thening follow the steps S3, otherwise, type of subject candidate item is generated according to the dispute type and demand and is selected for user, is used
Family selects type of subject from candidate item.
By working above, realizes identification dispute type, the purpose of demand and type of subject, be the meter of subsequent prediction model
It provides the foundation.
It is specifically described by taking following colloquial style description content as an example:
" my mother and what certain mother are uterine sisters, certain two people system cousin's relationship of I and what.2003
Just, we set up family, and after two months, our concealments are the identity of cousin, and in Yongxing County, Department of Civil Affairs is married at a registry.Now we
Because insensibility is basic, start long-term separation."
The consulting intention assessment result that intention assessment engine predicts above-mentioned colloquial style description content are as follows: dispute type: " wedding
Relation by marriage family ";Demand: " annulment of marriage ", " divorce ";Demand main body: " marital relations party ".
According to description content, the recognition result confidence level about dispute type and dispute main body is higher, so directly giving
As a result, and " annulment of marriage " and " divorce " confidence level are close under " marriage and family " this dispute type, so by intention assessment
The demand " annulment of marriage " and " divorce " that engine identifies are selected as candidate item for user.
Substantially, it is intended that identification engine is that the model based on text classification by inputting a Duan Wenben predicts one
The classification that a label, that is, user is intended to.
Further, as shown in figure 4, in step s3, the detection whether complete method of feature includes:
S31. it is intended to determine all characteristic informations to be obtained according to the consulting;
S32. the colloquial style merit is handled using feature calculation engine to describe, and look in colloquial style merit description
To the feature for being matched with characteristic information;
S33. judge with the presence or absence of non-matching characteristic in all characteristic informations, if so, testing result is characterized content not
Completely;
S34. all non-matching characteristics are ranked up according to feature importance;
S35. the ranking results are made into candidate problem list, and prompt user to supplement phase in a manner of puing question to user
The feature answered.
Specifically, in step S31, it is intended to determine all characteristic informations to be obtained, specifically, root according to the consulting
All characteristic informations for needing to obtain are determined according to specific demand.Continue by taking above-mentioned case as an example: it is assumed that user's final choice
" annulment of marriage ", for " annulment of marriage " this demand, it would be desirable to following characteristic information: " the annulment of marriage origin of an incident " and " marriage
Relationship ", " the annulment of marriage origin of an incident " and " marital relations " are exactly all characteristic informations that " annulment of marriage " this demand needs to obtain.
Preferably, it is predefined simultaneously for required characteristic information under various demands and characteristic information by legal staff
Storage.
As shown in figure 5, being retouched using feature calculation engine characteristic information feature according to required for demand to colloquial style merit
It states and carries out that comprehensive knowledge map is recalled, synonym extension, statistical nature, service feature are matched and handled, with from colloquial style case
The feature of characteristic information required for being matched with demand is found in feelings description.Or by taking above-mentioned case as an example: needing to match
Characteristic information be " the annulment of marriage origin of an incident " and " marital relations ", colloquial style merit description in " Yongxing County Department of Civil Affairs registration
Marriage ", " certain two people system cousin's relationship of I and what " are respectively the matching content of " the annulment of marriage origin of an incident " and " marital relations ",
That is all traffic issues-characteristic informations have been matched to from colloquial style description, so user does not need to carry out problem again
Supplement.If only having matched wherein some characteristic information, such as " marital relations ", then be not matched to characteristic information " marriage without
The corresponding question and answer of the effect origin of an incident ", which can be triggered, waits user to answer.Exception is, if current user information can be bright
True predicting is supported or is not supported, then problem still will not be triggered.
The form of supplement can be putd question in a manner of multiple-choice question, corresponding characteristic information can also be shown to prompt to user
User supplements corresponding feature for characteristic information, if such as " the annulment of marriage origin of an incident " this characteristic information in above-mentioned case
It is not matched, then being returned to " please supplement the annulment of marriage origin of an incident ", subsequent user inputs wedding in a manner of voice or text again
Then the relevant content of the invalid origin of an incident of relation by marriage is extracted " the annulment of marriage origin of an incident " matching by feature calculation engine in the text of supplement again
Feature.
According to the feature that feature calculation engine obtains, customer problem is targetedly inquired, will inquire customer problem
Only as the form of information supplement, rather than main form.Fusion Model based on structured features and text feature, such as
Wide&deep model combines the unstructured feature of structured features and textual form, combines natural language form
Unstructured information-feature, and manually comb dispute type, corresponding structured message-characteristic information below demand,
Solve the Sparse Problems of pure question and answer, the human knowledge of business expert cannot be efficiently used by also avoiding plain text feature.
Further, as shown in fig. 6, step S4 is specifically included:
S41. the prediction model is directed to demand according to all characteristic informations and the corresponding feature output of characteristic information
It wins a lawsuit the probability value of probability;
S42. obtain merit and predict reason template, and according to all characteristic informations and the corresponding feature of characteristic information and
Consulting result of the probability value output comprising prediction result and prediction reason.
Seeking advice from result is filling prediction reason and the resulting result of prediction result in prediction reason template.Prediction result is
One measurable number, rather than an ambiguous description, by this measurable probability value, user will be seen that civil
The prediction tendentiousness of case model, so that user has an intuitive impression;Prediction reason is based on user's input, characteristic matching
Situation and laws and regulations obtain.
Continue by taking above-mentioned case as an example, the probability of prediction model prediction is 93%, and corresponding grade is: must strive for.From telling
A possibility that a possibility that asking support, which is divided into, to attempt, can strive for, must strive for, more subsequent grade is supported is higher.
It seeks advice from comprising prediction result and prediction reason in result, and will be exported in prediction result and prediction reason filling template
Prediction is completed to user.The prediction result of above-mentioned case is 93%, must be striven for;Predict that reason is as follows:
Because you belong to the situation of [Department of Civil Affairs is married at a registry in Yongxing County], you and other side there are marital relations/
De facto marital relation.It is the invalid premise of marital relations there are marital relations.In reality, in fact it could happen that registration error, wrong card are stepped on
Note, the situation registered by means of name.These do not affect the establishment of de facto marital relation, and it is invalid not need to adjudicate this section of marital relations,
Most of solution routes are all the modes for taking administrative remedy, by registration authority's voluntarily error correction, cancel mistake registration.
" marriage law " Article 10: " there is one of following state, annulment of marriage: (one) bigamy;(2) have and forbid marriage
Kinship;(3) pre-marital with the disease for medically thinking not getting married, it not yet cures after marriage;(4) it does not arrive legal
Marriageable age." according to your description [certain two people system cousin's relationship of I and what], there are the annulment of marriage origins of an incident for you, therefore you want
Ask the request of confirmation annulment of marriage with lawfully recognized evidence.Division is determined outside the situation of annulment of marriage, it is also possible to there is " fraud " marriage.Dai Ling
Marriage certificate, it should be cancelled according to aggrieved party's will;Marriage certificate is falsely claimed as one's own, it is invalid to confirm, notice administration is removed
Pin registration;Wedding is deceived, except annulment of marriage is assert, notice administration will also notify judicial authority, investigate correlation cancel in accordance with the law outside
The criminal responsibility of personnel.
The present embodiment in view of the shortcomings of the prior art, firstly, being inputted based on user, is realized on the basis of intention assessment and is asked
Inscribe the automatic shunt of type, the corresponding problem types of automatic identification and type of subject;Secondly, the mouth inputted using text and voice
Mode of the languageization in conjunction with question and answer, the description and selection merit information of great convenience user;Secondly, the result of merit prediction is established
On the basis of the study of the data of magnanimity judgement document, the mode of data-driven is turned to from regular drive, covers the civil whole case
By the case data of ten million number of stages utilized, system has efficient scalability, and the powerful mould based on deep learning
Type automatically learns to inherent law therein and hiding relationship from merit and judgement, has good intelligent level;Again
It is secondary, by the form of pure question and answer be changed into based on colloquial style description based on, a small amount of problem supplement form and combination, reduce user need
Quantity, promotion user experience and the completion rate to be answered a question.Finally, the prediction conclusion of case, is capable of the prediction of quantitative,
So that user has an intuitive impression.
Embodiment two
As shown in fig. 7, present embodiment discloses a kind of descriptions of combination colloquial style and the intelligent civil case of question and answer to predict system
System, including preprocessing module 1, prediction module 2 and output module 3, wherein
Preprocessing module 1, whether comprising complete feature in the colloquial style merit description for detecting user's input,
If imperfect, user is prompted to supplement;
Prediction module 2, for calling prediction model and being won a lawsuit probability based on the output demand of complete feature;
Output module 3, for being predicted according to the prediction result and complete feature of the prediction module 2 based on merit
Reason template output consulting result.
Further, preprocessing module includes intention assessment engine 11, feature identification engine 12 and problem complementary module 13,
Wherein,
Intention assessment engine 11, the colloquial style merit for being inputted according to user describe to determine that the consulting of user is intended to;
Feature identifies engine 12, all characteristic informations to be obtained is determined for being intended to according to the consulting, in colloquial style
The feature for matching the characteristic information is found in merit description;
Problem complementary module 13, for non-matching characteristic to be made into candidate problem to prompt user to supplement in corresponding feature
Hold.
This system establishes the interface between Chinese judgement document's net, or establishes one and be stored with a large amount of judgement document's numbers
According to the memory space with law article regulation, access is called for prediction module and preprocessing module and output module.
Specific embodiment described herein is only an example for the spirit of the invention.The neck of technology belonging to the present invention
The technical staff in domain can make various modifications or additions to the described embodiments or replace by a similar method
In generation, however, it does not deviate from the spirit of the invention or beyond the scope of the appended claims.
Although preprocessing module, prediction module, output module, intention assessment engine and feature is used more herein to know
The terms such as other engine, but it does not exclude the possibility of using other terms.The use of these items is only for more easily retouch
State and explain essence of the invention;Being construed as any additional limitation is disagreed with spirit of that invention.
Claims (10)
1. a kind of intelligent civil case prediction technique of combination colloquial style description and question and answer, which comprises the following steps:
S1. the colloquial style merit description of user's input is received;
S2. describe to determine that the consulting of user is intended to according to the colloquial style merit;
S3. whether complete detection feature is intended to according to the consulting, if so, thening follow the steps S4, otherwise user is prompted to mend
Fill corresponding feature;
S4. prediction model is called to export corresponding consulting result to user according to complete feature.
2. the intelligent civil case prediction technique of combination colloquial style description and question and answer according to claim 1, feature exist
In, in step sl, the colloquial style merit description includes that the merit description of literal type and the merit of sound-type describe,
And when being described for the merit of sound-type, first passes through the speech recognition system and convert speech into text.
3. the intelligent civil case prediction technique of combination colloquial style description and question and answer according to claim 2, feature exist
In in step s 2, the consulting is intended to include problem types and dispute main body, and described problem type includes dispute type and tells
It asks.
4. the intelligent civil case prediction technique of combination colloquial style description and question and answer according to claim 3, feature exist
In in step s 2, determining consulting is intended in the following manner:
S21. dispute type and demand are identified using intention assessment engine, and judges whether the dispute type and demand meet
Confidence level requirement, if so, thening follow the steps S22, the candidate item for otherwise generating dispute type and demand type is selected for user;
S22. type of subject is identified using intention assessment engine and judge whether the type of subject meets confidence level requirement, if
It is to then follow the steps S3, otherwise, type of subject candidate item is generated according to the dispute type and demand and is selected for user.
5. the intelligent civil case prediction technique of combination colloquial style description and question and answer according to claim 4, feature exist
In in step s3, the detection whether complete method of feature includes:
S31. it is intended to determine all characteristic informations to be obtained according to the consulting;
S32. the colloquial style merit is handled using feature calculation engine to describe, and find in colloquial style merit description
Feature with the characteristic information;
S33. judge in all characteristic informations with the presence or absence of non-matching characteristic, if so, to be characterized content imperfect for testing result.
6. the intelligent civil case prediction technique of combination colloquial style description and question and answer according to claim 5, feature exist
In in step s3, prompt user supplements corresponding feature by the following method:
S34. all non-matching characteristics are ranked up according to feature importance;
S35. the ranking results are made into candidate problem list, and prompt user's supplement corresponding in a manner of puing question to user
Feature.
7. the intelligent civil case prediction technique of combination colloquial style description and question and answer according to claim 6, feature exist
In step S4 is specifically included:
S41. the prediction model is won a lawsuit the probability value of probability according to the output of all features for demand;
S42. it obtains merit and predicts reason template, and include prediction result and prediction according to all features and probability value output
The consulting result of reason.
8. the intelligent civil case prediction technique of combination colloquial style description and question and answer according to claim 7, feature exist
In the prediction model is trained acquisition to deep learning model by using judgement document's data by mark.
9. a kind of intelligent civil case forecasting system of combination colloquial style description and question and answer, which is characterized in that including pre-processing mould
Block (1), prediction module (2) and output module (3), wherein
Preprocessing module (1), for whether detecting in the colloquial style merit description that user inputs comprising complete feature, if
It is imperfect, then prompt user to supplement;
Prediction module (2), for calling prediction model and being won a lawsuit probability based on the output demand of complete feature;
Output module (3), for being predicted according to the prediction result and complete feature of the prediction module (2) based on merit
Reason template output consulting result.
10. the intelligent civil case forecasting system of combination colloquial style description and question and answer according to claim 9, feature exist
In the preprocessing module (1) includes intention assessment engine (11), feature identification engine (12) and problem complementary module
(13), wherein
Intention assessment engine (11), the colloquial style merit for being inputted according to user describe to determine that the consulting of user is intended to;
Feature identifies engine (12), all characteristic informations to be obtained is determined for being intended to according to the consulting, in colloquial style case
The feature for matching the characteristic information is found in feelings description;
Problem complementary module (13), for non-matching characteristic to be made into candidate problem to prompt user to supplement in corresponding feature
Hold.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910348700.2A CN110222866A (en) | 2019-04-28 | 2019-04-28 | In conjunction with the intelligent civil case forecasting system and method for colloquial style description and question and answer |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910348700.2A CN110222866A (en) | 2019-04-28 | 2019-04-28 | In conjunction with the intelligent civil case forecasting system and method for colloquial style description and question and answer |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110222866A true CN110222866A (en) | 2019-09-10 |
Family
ID=67820131
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910348700.2A Pending CN110222866A (en) | 2019-04-28 | 2019-04-28 | In conjunction with the intelligent civil case forecasting system and method for colloquial style description and question and answer |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110222866A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111368032A (en) * | 2020-02-29 | 2020-07-03 | 重庆百事得大牛机器人有限公司 | Daily language identification method for legal consultation |
CN111368098A (en) * | 2020-03-31 | 2020-07-03 | 杭州实在智能科技有限公司 | Scenario-based legal consultation assessment system |
CN111597307A (en) * | 2020-05-18 | 2020-08-28 | 山西大学 | Judicial judgment reasoning method based on interpretable causal model |
CN113886555A (en) * | 2021-12-03 | 2022-01-04 | 北京大学 | Legal consultation service method and device |
CN116385216A (en) * | 2023-04-11 | 2023-07-04 | 南北联合信息科技有限公司 | Intelligent civil case prediction system combining spoken language description and question answering |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106095872A (en) * | 2016-06-07 | 2016-11-09 | 北京高地信息技术有限公司 | Answer sort method and device for Intelligent Answer System |
CN106296495A (en) * | 2016-08-09 | 2017-01-04 | 点击律(上海)网络科技有限公司 | The Forecasting Methodology of a kind of lawsuit result and system |
CN108549656A (en) * | 2018-03-09 | 2018-09-18 | 北京百度网讯科技有限公司 | Sentence analytic method, device, computer equipment and readable medium |
-
2019
- 2019-04-28 CN CN201910348700.2A patent/CN110222866A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106095872A (en) * | 2016-06-07 | 2016-11-09 | 北京高地信息技术有限公司 | Answer sort method and device for Intelligent Answer System |
CN106296495A (en) * | 2016-08-09 | 2017-01-04 | 点击律(上海)网络科技有限公司 | The Forecasting Methodology of a kind of lawsuit result and system |
CN108549656A (en) * | 2018-03-09 | 2018-09-18 | 北京百度网讯科技有限公司 | Sentence analytic method, device, computer equipment and readable medium |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111368032A (en) * | 2020-02-29 | 2020-07-03 | 重庆百事得大牛机器人有限公司 | Daily language identification method for legal consultation |
CN111368032B (en) * | 2020-02-29 | 2020-12-11 | 重庆百事得大牛机器人有限公司 | Daily language identification method for legal consultation |
CN111368098A (en) * | 2020-03-31 | 2020-07-03 | 杭州实在智能科技有限公司 | Scenario-based legal consultation assessment system |
CN111368098B (en) * | 2020-03-31 | 2023-04-07 | 杭州实在智能科技有限公司 | Scenario-based legal consultation assessment system |
CN111597307A (en) * | 2020-05-18 | 2020-08-28 | 山西大学 | Judicial judgment reasoning method based on interpretable causal model |
CN113886555A (en) * | 2021-12-03 | 2022-01-04 | 北京大学 | Legal consultation service method and device |
CN116385216A (en) * | 2023-04-11 | 2023-07-04 | 南北联合信息科技有限公司 | Intelligent civil case prediction system combining spoken language description and question answering |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110222866A (en) | In conjunction with the intelligent civil case forecasting system and method for colloquial style description and question and answer | |
Tien et al. | Subjective well-being in tourism research | |
CN111026842B (en) | Natural language processing method, natural language processing device and intelligent question-answering system | |
Saini et al. | Systematic synthesis of qualitative research | |
CN104794768B (en) | Attendance checking system and Work attendance method that timing automatic operation work attendance is calculated | |
CN103208092B (en) | The system of consultation information is provided based on internet | |
LU100314B1 (en) | Method and system for predicting academic achievements of students based on naive bayesian model | |
CN110222145A (en) | A kind of intelligence law appraisal procedure and system | |
CN110232447B (en) | Deep reasoning method for legal case | |
CN113744872A (en) | Intelligent interactive Alzheimer's disease auxiliary screening system | |
Jardine et al. | Valuing desistence? A social return on investment case study of a throughcare project for short-term prisoners | |
CN110414837A (en) | Based on mistake because of the man-machine interactive system of analysis | |
CN111696648A (en) | Psychological consultation platform based on Internet | |
Ferreira et al. | Multimodal and adaptable medication assistant for the elderly: A prototype for interaction and usability in smartphones | |
Tebaldi et al. | The power of the president: a quantitative narrative analysis of the diary of an Italian head of state (2006–2013) | |
CN113420059A (en) | Method and device for actively treating citizen hot line problem | |
CN116955445A (en) | Complaint event data mining analysis method and system based on information extraction | |
Al Reshidi | The impact of a training intervention on emotional intelligence, leadership styles, self-efficacy and perception of sense of power in a university nursing faculty in Saudi Arabia | |
Papakonstantinou et al. | A bone of contention: teacher evaluation system in Greece | |
Nylund | Comparing the efficiency and quality of civil justice in Scandinavia: the role of structural differences and definitions of quality | |
Tshabalala | The role of community participation in the integrated development plan of Govan Mbeki Municipality | |
Joungtrakul | Thai engineers’ readiness to cope with the free flow of skilled labor in the ASEAN Economic Community | |
CN115394431A (en) | Intelligent medical service system | |
Webber et al. | Applying research evidence in social work practice: Seeing beyond paradigms | |
KR20120011107A (en) | System and method for diagnosing learning indicator of language area |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190910 |