CN109360127A - A kind of chain of evidence relational graph modeling method - Google Patents
A kind of chain of evidence relational graph modeling method Download PDFInfo
- Publication number
- CN109360127A CN109360127A CN201811272642.1A CN201811272642A CN109360127A CN 109360127 A CN109360127 A CN 109360127A CN 201811272642 A CN201811272642 A CN 201811272642A CN 109360127 A CN109360127 A CN 109360127A
- Authority
- CN
- China
- Prior art keywords
- evidence
- chain
- tie
- text
- point
- 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 29
- 230000000007 visual effect Effects 0.000 claims abstract description 7
- 238000005516 engineering process Methods 0.000 claims abstract description 6
- 239000000284 extract Substances 0.000 claims description 19
- 238000000605 extraction Methods 0.000 claims description 18
- 230000014509 gene expression Effects 0.000 claims description 13
- 230000011218 segmentation Effects 0.000 claims description 6
- 238000011524 similarity measure Methods 0.000 claims description 3
- 238000012800 visualization Methods 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 239000000463 material Substances 0.000 description 5
- 238000013461 design Methods 0.000 description 3
- 230000002045 lasting effect Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 210000000056 organ Anatomy 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000000739 chaotic effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 230000008021 deposition Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000010422 painting Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training 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
- G06Q50/18—Legal services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Technology Law (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Geometry (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of chain of evidence relational graph modeling methods, belong to the application technology that visual modeling method is combined with law court's trial business.Chain of evidence relational graph modeling method includes the following steps: to be loaded into and decompose both sides' evidence list, calculate chain head short text, cross-examination with accept and believe, it is loaded into and decomposes true text, calculate tie-point short text, it calculates the similitude of chain head short text and tie-point short text and establishes mutual connection relationship, chain of evidence relational graph is drawn, the XML format of chain of evidence relational graph is exported.Method of the invention can assist judge when hearing a case, and comb evidence by visual modeling tool and construct chain of evidence, show the evidence and evidence linking relationship of some case in a more intuitive way.
Description
Technical field
The invention discloses a kind of chain of evidence relational graph modeling method, belongs to visual modeling method and law court tries business
The application technology combined.
Background technique
The writing of judgement document is the basic training of judge, and argues, and is the foundation stone that judge writes judgement document.It cuts
Realification solution contradiction and disputes show that the administration of justice is efficient and just, establish judicial authority and public affairs are believed, just must be raising judgement document's
It argues horizontal as a groundwork.Strive allowing every judgement document, can accomplish to argue thoroughly sufficiently, with the case that concludes.
Judgement document argues or shows the fairness of administration of justice, evaluates the quality of judgement document and the major criterion of level, is hair
An important factor for waving judgement document's multi-functional.
It improves judgement document's mass and has become the significant obstacle for improving trial quality and promoting judicial authority, exist serious
Influence the realization of trial open, justice.In the case where social contradications are in the majority instantly, a large amount of social contradications pour in law court,
And the finiteness of judicial resources itself, cause judge to face serious operating pressure, a large amount of case up for trial judge,
Hard work pressure causes judgement document generally existing in current juridical practice to prove insufficient, the hypodynamic drawback of persuasion,
Make party in face of verdicts of court simultaneously, does not understand law court is how to assert the fact, and judged according to what, into
And it is made to throw doubt upon fairness, the legitimacy of judge's result, appeal is directly resulted in, appeals, twine phenomena such as telling, therefore is added
Strong judgement document has its realistic meaning.Although many law courts propose many requirements, establish to improve judgement document's mass
Relevant working mechanism is perfected, so that the level of arguing of judgement document increases.But in terms of the whole country, judgement document
The case where thinking little of arguing, will not arguing still remains, or even has judgement document full of mistakes.
To solve the above-mentioned problems, the assessment of arguing of document, such as TF- are completed by the intelligentized mode of text analyzing
IDF extracts keyword and LSA latent semantic analysis, the great principle that judgement document can be avoided to argue to a certain extent
Mistake substantially reduces the workload of judge, reduces the fault of the relevant technologies or non-principle, self-consistent when accomplishing to argue,
Great convenience is brought to administration of justice.
Another level, in order to avoid the unfairness of the judicial adjudication, the mode analyzed using intelligent text, to detect one
Whether rationally arguing for document, ensure that the fairness of trial to a certain extent.And when judge decides a case, root
It can be that judge pushes law article with law article recommender system, judge can join when deciding a case when according to case reference entity method
Examine the law article of recommendation.
So the intelligentized mode of text analyzing carry out judgement document argue assessment be to have very much realistic meaning.
Summary of the invention
Judge always faces various problems when handling actual case, in the tool that none is specially designed
In the case of, it if the case encountered is more complicated, will lead to the problem of a series of: (1) in case, work as from each
The evidence quantity of the sources of evidence such as thing people, detection organ is big and complicated, it is difficult to index；(2) key message in evidence is too
It is more, clearly incidence relation is difficult to find that before overall contrast；(3) after deducing possible fact of case, if it
The evidence availability or content of preceding submission are changed, and modify factural information associated with it is also very troublesome one
Part thing.
The problem to be solved in the present invention designs a kind of general chain of evidence relationship modeling aiming at above-mentioned business demand
Language is for describing the process that chain of evidence is argued；A set of patterned chain of evidence is developed on the basis of this language simultaneously
Relationship modeling tool provides semi-automation, business function mathematics library tool abundant, and judge is facilitated to manage evidence letter
Breath visually shows the relationship between evidence and the reasoning fact, improves judge and arranges case, writes judgement document's
Efficiency.
The technical solution of the present invention is as follows:
1. belonging to visual modeling method the invention discloses a kind of chain of evidence relational graph modeling method and law court trying industry
The application technology that business combines.Chain of evidence relational graph modeling method includes the following steps (as shown in Figure 1):
Step (1) typing and the plaintiff's evidence (or accusing party's evidence) and defendant's evidence text for decomposing case obtain double
Square evidence list allows user to correct and edit the text of the evidence list after decomposing.
Step (2) decomposes obtained both sides' evidence list according to step (1), carries out to the evidence text of evidence list crucial
Element extracts, and key element extraction is exactly the chain head short text for calculating each evidence, that is, obtains the several of each evidence serobila
A chain head short text, and several chain head short texts are listed in below associated evidence and are sequentially shown, and allow user into
Row amendment and editor's chain head short text.
Step (3) carries out cross-examination and accepts and believe operation, evidence list and its chain head short text after obtaining cross-examination and accepting and believing.
Step (4) be loaded into and decompose true text (fact text may be from indictment accusation content or
The fact that be civil bill of complaint description), allow user to correct and edit the text of the true list after decomposing；And to true text
This progress key element extraction, key element extraction is exactly the tie-point short text for calculating each fact, that is, obtains each thing
Several real tie-point short texts, and several tie-point short texts are listed in associated true lower section and are sequentially shown,
And user is allowed to be modified and edit tie-point short text.
Step (5) calculates the similitude of chain head short text and tie-point short text, and determines the connection of chain head and tie-point
Relationship.The Similarity measures are measured using COS distance, and it is 0.9 that threshold value, which is arranged, i.e. similarity is more than or equal to 0.9, just
Determine that chain head short text is related to tie-point short text, and establishes mutual connection relationship.
Step (6) according to chain head and tie-point connection relationship, draw chain of evidence relational graph, the figure include evidence serobila,
Chain head, tie-point and true node, and its between connection relationship.
The XML format of step (7) offer chain of evidence relational graph.
2. the entity in chain of evidence relational model, the definition of relationship and rule are as follows:
(1) entity in relational model is divided into four classes, is respectively: evidence serobila, chain head, tie-point and the fact, wherein demonstrate,proving
Single evidence is described according to serobila, chain head describes the key message of evidence, and tie-point describes mutually to print in evidence
The fact that the key message of card is embodied text, in the known integrated circuit it is a fact that the either civil bill of complaint of the accusation content from indictment
The fact describe.
(2) connection relationship in relational model includes three kinds: the connection relationship of " evidence serobila-chain head ", the relationship is by step
Suddenly the calculated result of (2) is established；The connection relationship of " chain head-tie-point ", the relationship are established by the checkout result of step (4)；
The connection relationship of " tie-point-fact ", the result are established by the checkout result of step (5).The connection of " evidence serobila-chain head "
Relationship description is inclusion relation between evidence itself and evidence key message, the connection relationship description of " chain head-tie-point "
Be evidence key message it is independent or partly confirm a fact and the verifying relationship that generates, " tie-point-fact "
Connection relationship describes the inclusion relation between true associated key message itself.
(3) rule in relational model are as follows: evidence serobila can only be connected with chain head, and an evidence serobila can produce more
A chain head, a chain head can only correspond to an evidence serobila；One chain head can be connected with multiple tie-points, and tie-point can
To connect with multiple chain heads, a true node can be contacted with multiple tie-points.
3. evident information typing according to claim 1 and dividing method, it is characterised in that will in step (1)
Original evidence text segmentation with multiple paragraphs is sentenced at several independent evident informations for what an evidence description terminated
Calibrating standard includes: to encounter newline, encounters fullstop, and has specific Chinese character serial number or the serial number of Arabic numerals to indicate
As the segmentation boundary etc. between each card text description.
4. evidence text elements according to claim 1 extract mode, it is characterized in that according to 4W1H in step (2)
Key element extracts strategy, obtains the keyword set of each evidence serobila and each fact, specifically includes:
Step (2.1) segments true and evident information, and anolytic sentence dependency structure relationship, and auxiliary uses canonical
Expression formula extracts key element What, i.e., the things being related in information；
Step (2.2) extracts key element When using regular expression from true and evident information, i.e., relates in information
And the time arrived；
Step (2.3) segments true and evident information, and analyzes part of speech, phrase structure relationship, and extraction is critical to
Plain Where, i.e., the place being related in information；
Step (2.4) segments true and evident information, and analyzes part of speech, extracts key element Who, i.e., in information
The party being related to；
Step (2.5) extracts key element How much, i.e. information using regular expression from true and evident information
In the quantity that is related to, mainly include the amount of money and weight etc..
5. patterned modeling tool according to claim 1, chain of evidence relational graph is drawn in modeling tool
Three kinds of connection relationships, three connection relationships of chain of evidence relational graph are respectively by the calculating of step (2), step (4) and step (5)
As a result it establishes, provides complete graphical tool to support model definition and the rule in right 2.
6. the chain of evidence relational graph according to claim 1, drawn with chain of evidence relational graph modeling tool, in addition to visualization is schemed
Outside the storage form of shape, there are also XML storage formats, to support the chain of evidence relational model of the case in different terminals computer
It is loaded into and edits, conveniently handle a case.
7. mathematics library tool used in step (6) is the chain of evidence relationship visualization modeling work specially developed
Tool, function mainly includes the following:
(1) it adds, delete, editor's pel, the type of pel includes four kinds: evidence serobila (blue side length is rectangular), chain head (circle
Shape), tie-point (square), true node (orange side length is rectangular)；
(2) connection relationship is constructed between associated pel；As mentioned in the text, the relation in evidence chain model
It is three kinds, is relationship between relationship, tie-point and the chain head between evidence serobila and chain head and tie-point and the fact respectively
Relationship between node, the relationship between evidence serobila and chain head are one-to-many relationships, the relationship between tie-point and chain head
It is one-to-many relationship, the relationship between true node and tie-point is one-to-many relationship；
(3) auto arrangement pel；After random addition enters several pels, tool offer can be with Automatic Typesetting
Function seems it neatly all pels typesetting according to certain rules before, beautiful；
(4) it is directed to the grammer of the chain of evidence relational graph, designs XML Schema and specific XML storage format, and can
To use the chain of evidence relational graph modeling tool operation such as to be read out, edit, save.
Detailed description of the invention
Fig. 1 chain of evidence relational graph modeling process flow chart
The graphical examples of Fig. 2 chain of evidence relation model figure
The original evidence of Fig. 3 chain of evidence modeling tool decomposes panel
Newly-built pel part in the visualization model of Fig. 4 chain of evidence modeling tool
Three kinds of connection relationships of visualization model in Fig. 5 chain of evidence relational graph modeling tool
Layout type in Fig. 6 chain of evidence relational graph modeling tool after the automation typesetting of visualization model
The XML format example of Fig. 7 chain of evidence relational graph modeling tool storage
Specific embodiment
To be more clear the object, technical solutions and advantages of the present invention, below in conjunction with attached drawing and specific example to this
Invention is described in detail.
The present invention is intended to provide the visual modeling tool of a set of evidence chain relation needs to design complete before this
Chain of evidence modeling language is closed to define and provide the connection relationship between entity and entity in chain of evidence relational model
System is as shown in Fig. 2, specific as follows:
(1) entity in relational model is divided into four classes, is respectively: evidence serobila, chain head, tie-point and the fact, wherein demonstrate,proving
Single evidence is described according to serobila, chain head describes the key message of evidence, and tie-point describes mutually to print in evidence
The fact that the key message of card is embodied text, in the known integrated circuit it is a fact that the either civil bill of complaint of the accusation content from indictment
The fact describe.
(2) connection relationship in relational model includes three kinds: the connection relationship of " evidence serobila-chain head ", " chain head-connection
Connection relationship, the connection relationship of " tie-point-fact " of point "；The connection relationship of " evidence serobila-chain head " describes evidence
Inclusion relation between evidence key message in itself, the connection relationship description of " chain head-tie-point " is the key that evidence letter
The verifying relationship for ceasing independence or partly one fact of confirmation and generating, the connection relationship of " tie-point-fact " describe
Inclusion relation between true associated key message itself.
(3) rule in relational model are as follows: evidence serobila can only be connected with chain head, and an evidence serobila can produce more
A chain head, a chain head can only correspond to an evidence serobila；One chain head can be connected with multiple tie-points, and tie-point can
To connect with multiple chain heads, a true node can be contacted with multiple tie-points.
According to designed rule and definition, chain of evidence relationship modeling tool is developed, steps are as follows for the use of the tool:
Step (1) typing and the plaintiff's evidence (or accusing party's evidence) and defendant's evidence text for decomposing case obtain double
Square evidence list allows user to correct and edit the text of the evidence list after decomposing.
Step (2) decomposes obtained both sides' evidence list according to step (1), carries out to the evidence text of evidence list crucial
Element extracts, and key element extraction is exactly the chain head short text for calculating each evidence, that is, obtains the several of each evidence serobila
A chain head short text, and several chain head short texts are listed in below associated evidence and are sequentially shown, and allow user into
Row amendment and editor's chain head short text.
Step (3) carries out cross-examination and accepts and believe operation, evidence list and its chain head short text after obtaining cross-examination and accepting and believing.
Step (4) be loaded into and decompose true text (fact text may be from indictment accusation content or
The fact that be civil bill of complaint description), allow user to correct and edit the text of the true list after decomposing；And to true text
This progress key element extraction, key element extraction is exactly the tie-point short text for calculating each fact, that is, obtains each thing
Several real tie-point short texts, and several tie-point short texts are listed in associated true lower section and are sequentially shown,
And user is allowed to be modified and edit tie-point short text.
Step (5) calculates the similitude of chain head short text and tie-point short text, and determines the connection of chain head and tie-point
Relationship.The Similarity measures are measured using COS distance, and it is 0.9 that threshold value, which is arranged, i.e. similarity is more than or equal to 0.9, just
Determine that chain head short text is related to tie-point short text, and establishes mutual connection relationship.
Step (6) according to chain head and tie-point connection relationship, draw chain of evidence relational graph, the figure include evidence serobila,
Chain head, tie-point and true node, and its between connection relationship.
The XML format of step (7) offer chain of evidence relational graph.
For convenience of explanation, the chain of evidence modeling process that true case will be simulated below, from obtaining the original card of case
It is believed that breath starts, complete chain of evidence Visualization Model is gradually obtained, overall flow figure is as shown in Figure 1.
1. after putting on record, which submits evidence, various relevant evidences have text description.In practical applications
Diversified evidence may be faced, no matter but the evidence of what form and what source, be required to additional one section about this
The text expression of evidence, therefore the present invention carries out all kinds of calculating and text analyzing based on the description of evidence text.
(1) step operation for original evidence description is to decompose evidence and classify to evidence, each works as thing
The evidence that people is submitted may be multistage, and in each section, party may be the elaboration carried out for a certain situation.To point
The benefit that text after cutting carries out next step operation is just that of avoiding the same evidence serobila and generates excessive chain head, so that knot
Structure is not clear enough.The category of evidence of criminal case includes: material evidence, documented evidence, testimony of witnesses, victim's statement, suspect,
Defendant makes a deposition and explanation, expert opinion, investigates on the spot, checks, recognizing, the notes such as investigative test, audiovisuals, electronic data.The people
The category of evidence of thing case includes: litigant's statement, documented evidence, material evidence, audiovisuals, testimony of witnesses, electronic data, identification knot
By records of inquests.Administrative evidence type includes: documented evidence, material evidence, audiovisual materials, testimony of witnesses, the statement of party,
Expert's conclusion, records of inquests and scene notes.
2. the purpose of step (2) is valuable key message in the extraction evidence for automation, the mode of extraction is adopted
With the mode of 4W1H (When, Where, What, Who, How much).The content of evidence obtained in previous step can be with
Be made of multiple words, and wherein many words be all it is useless, do not need to be included in calculating, it is therefore desirable to evident information
It carries out Text Pretreatment and extracts 4W1H key message, specific steps include:
Step (2.1) uses two methods of structural formula and syntactic analysis for the extraction of key message What.Structural formula is
Refer to that structural formula method refers to for the things with fixed structure, is extracted using regular expression, such as in punctuation marks used to enclose the title
File name.Syntactic analysis method is the extracting method for the not things of special identifier, first to true and evident information
It is segmented, and anolytic sentence dependency structure relationship, things would generally be as the subject or object of sentence, it is possible to according to
The interdependent syntactic structure analyzed, extracts subject and object, screens out the word for wherein belonging to title and place, then root
It is perfect that word is carried out according to surely middle relationship.Here using HanLP as participle and Sentence analysis tool；
Step (2.2) uses regular expression method for the extraction of key message When.Regular expression can match
Out from most accurate " when X X month X day X X points " to the literary style of a variety of temporal expressions such as rough " X "；
Step (2.3) uses two methods of morphological analysis and syntactic analysis for the extraction of key message Where.It is right first
True and evident information is segmented, and analyzes part of speech and phrase structure relationship.Continuous Chinese text is segmented almost
Necessary step in all natural language processing methods could be further processed text after participle.
Currently, Words partition system is the more mature system of a theory, use Ansj Chinese word segmentation machine as our participle system here
System segments mode as our participle mode using NLP.Ansj Chinese word segmentation machine can be to the part of speech of each word after participle
It is labeled, part of speech is the place extracted required for the word of S (place word) is.In view of the error rate of participle, it is also necessary to
The syntactic analysis based on preposition is carried out, because, by establishing preposition list, extracting after place typically occurs in preposition
Word after preposition is as place；Again because of the problem of being likely to occur excessive division when participle, also need to preposition it
Multiple words afterwards are judged, and form complete place by connecting into；
Step (2.4) uses part of speech analysis method for the extraction of key message Who.Because of the personage in evident information,
Not only comprising individual, it is also possible to unit or organ, so name cannot be extracted merely.Believe firstly the need of to true and evidence
Breath is segmented, and then analyzes the part of speech and phrase structure relationship of word, it (includes Chinese name, foreign language people that part of speech is started with NR
Name and transliteration name) and part of speech be the word of NT (group of mechanism name) be all the title of required extraction；
Step (2.5) uses regular expression method for the extraction of key message How much.Regular expression can be with
Match the type of the amount of money (such as " X yuans "), weight (such as " X grams ") a variety of quantity words；
Because the key element in information is likely to occur the case where repetition statement, also need to consider to go in above step
Weight problem, it is ensured that all without duplicate word in every kind of key element, the display effect of first two steps is shown in that (chain of evidence models work to Fig. 3
The original evidence of tool decomposes panel).
3. operation is exactly in mathematics library tool in next step after having obtained extracting the evident information of key element
On obtain the Visualization Model of evidence chain relation, the function following points of heretofore described mathematics library tool:
(1) it adds, delete, editor's pel.
Four kinds of entities in chain of evidence relational model have been described above, have been evidence serobila respectively, chain head, tie-point with
And true node.Graphical chain of evidence relational graph modeling tool realizes the addition of these four entities, editor, and deletes behaviour
Make；The mode of addition is divided into common button addition (pel is added on painting canvas after pressing corresponding button), and shortcut key adds
Add and (after the shortcut key on lower keyboard, i.e., add pel in mouse position), lasting addition switch be (point lower switch
After button, into lasting addition state, a pel can be created by clicking left mouse button every time) etc., display effect is shown in Fig. 4
(the newly-built pel part in visualization model).
In order to facilitate editor, also developed frame modeling block and right-click menu in this tool, the effect of frame modeling block be using
Family can select a certain range of pel, so as to the movement of batch, delete pel；The effect of right-click menu then mentions
For more intuitive option of operation, user can operation that directly right button is needed on needing the pel edited, such as create
Arrow, editor's pel, deletion pel, duplication pel etc..
(2) connection relationship is constructed between associated pel.
As mentioned in the text, the relationship in evidence chain model is divided into three kinds, is respectively: between evidence serobila and chain head
Connection relationship, tie-point between connection relationship, tie-point and chain head and the connection relationship between the fact.It should be noted that
Connection relationship between evidence serobila and chain head is to determine that when chain head generates, and can not delete, because of chain head
There must be the evidence serobila in its source, so not allowing the chain head occurred and any evidence serobila does not all connect；And tie-point
Connection relationship, tie-point and true connection relationship between chain head can only submitted then without excessively stringent limitation
When check, the content of inspection includes: whether to exist and all disjunct tie-point of any chain head, whether there is and any
The all disjunct true node of tie-point whether there is the tie-point being connected with more than a true node, because of tie-point, thing
The editor of physical node and connection are all by user's operation, and can be supplied to user more freely edits space in the process for this,
Display effect is shown in Fig. 5 (three kinds of connection relationships of visualization model in chain of evidence relational graph modeling tool).
(3) auto arrangement pel.
With the raising of the quantity and complexity of evidence involved in case, chain of evidence illustraton of model needed for analyzing case
Size can also become huge, and composition can become mixed and disorderly；In order to keep the readability of chain of evidence figure stronger, this tool is provided
The primitive rule of the function of pel auto arrangement, the function is: four class pels is individually placed in four columns, such as Fig. 2, each
Evidence serobila can all be connected with several chain heads, form tree, and have close connected rule, chain head and tie-point
Between connection, the connection between tie-point and true node may then go out without close connected rule, therefore theoretically
Existing very complicated structure, in order to solve the chaotic situation being likely to occur, in this tool can also to the node currently selected and its
The node being connected carries out highlighted processing, and preferable readable to guarantee to have, display effect is shown in that Fig. 6 (build by chain of evidence relational graph
Layout type in die worker's tool after the automation typesetting of visualization model).
(4) there is specific XML format, and can be read out and save.For the grammer of the chain of evidence relational graph,
XML Schema and specific XML storage format are designed, and the chain of evidence relational graph modeling tool can be used and read
The operation such as take, edit, saving.Dedicated XML schema has also been devised in this tool in the characteristics of for evidence chain model, uses
In describing and storing evidence chain model, (chain of evidence relational graph modeling tool stores the topology example of the schema as shown in Figure 7
XML format example).
Above is exactly to retouch to chain of evidence modeling language included in the present invention and the detailed of corresponding modeling tool
It states.The present invention has the advantages that as follows: modeling to the judicial middle process argued using chain of evidence of deciding a case, so that user exists
Evidence material can be combed in a manner of visual modeling according to chain of evidence relational model during being decided a case.The present invention
Mating to develop a chain of evidence relational graph modeling tool, which can be with the key message in assisted extraction evidence, while energy
Clearly chain of evidence relational graph arranges evident information for judge and sentences with advantage easy to operate, clear in structure for enough buildings
Case thinking provides effective auxiliary tool, argues and analyzes for judgement document, the verdict of case, judgement document write work
Make to provide and helps and prepare.
It needs to define, the invention is not limited to specific configuration described above and shown in figure and processing.Also,
For brevity, the detailed description to known method technology is omitted here.Current embodiment is all counted as in all respects
It is exemplary rather than limited, the scope of the present invention is by appended claims rather than foregoing description defines, and falls into power
Whole changes in the range of meaning and equivalent that benefit requires are to all be included among the scope of the present invention.
Claims (6)
1. belonging to visual modeling method the invention discloses a kind of chain of evidence relational graph modeling method with law court and trying business phase
In conjunction with application technology.Chain of evidence relational graph modeling method includes the following steps:
Step (1) typing and the plaintiff's evidence (or accusing party's evidence) and defendant's evidence text for decomposing case obtain both sides' card
According to list, user is allowed to correct and edit the text of the evidence list after decomposing.
Step (2) decomposes obtained both sides' evidence list according to step (1), carries out key element to the evidence text of evidence list
It extracts, key element extraction is exactly the chain head short text for calculating each evidence, that is, obtains several chains of each evidence serobila
Head short text, and several chain head short texts are listed in below associated evidence and are sequentially shown, and user is allowed to repair
Just with editor chain head short text.
Step (3) carries out cross-examination and accepts and believe operation, evidence list and its chain head short text after obtaining cross-examination and accepting and believing.
Step (4) be loaded into and decompose true text (the fact text may be from indictment accusation content or the people
The fact that the thing bill of complaint, describes), allow user to correct and edit the text of the true list after decomposing；And true text is carried out
Key element extracts, and key element extraction is exactly the tie-point short text for calculating each fact, if obtaining each fact
Dry tie-point short text, and several tie-point short texts are listed in associated true lower section and are sequentially shown, and are allowed
User is modified and edits tie-point short text.
Step (5) calculates the similitude of chain head short text and tie-point short text, and determines the connection relationship of chain head and tie-point.
The Similarity measures are measured using COS distance, and it is 0.9 that threshold value, which is arranged, i.e. similarity is more than or equal to 0.9, decides that chain
Head short text is related to tie-point short text, and establishes mutual connection relationship.
Step (6) according to chain head and tie-point connection relationship, draw chain of evidence relational graph, the figure include evidence serobila, chain head,
Tie-point and true node, and its between connection relationship.
The XML format of step (7) offer chain of evidence relational graph.
2. the entity in chain of evidence relational model, the definition of relationship and rule are as follows:
(1) entity in relational model is divided into four classes, is respectively: evidence serobila, chain head, tie-point and the fact, wherein chain of evidence
Body describes single evidence, and chain head describes the key message of evidence, what tie-point described mutually to confirm in evidence
The fact that key message is embodied text, in the known integrated circuit it is a fact that the accusation content from indictment either the civil bill of complaint the fact
Description.
(2) connection relationship in relational model includes three kinds: the connection relationship of " evidence serobila-chain head ", the relationship is by step (2)
Calculated result establish；The connection relationship of " chain head-tie-point ", the relationship are established by the checkout result of step (4)；" tie-point-
The connection relationship of the fact ", the result are established by the checkout result of step (5).The connection relationship description of " evidence serobila-chain head "
It is the inclusion relation between evidence itself and evidence key message, the connection relationship of " chain head-tie-point " describes evidence
Key message is independent or partly confirms a fact and the verifying relationship that generates, and the connection relationship of " tie-point-fact " is retouched
What is stated is the inclusion relation between true associated key message itself.
(3) rule in relational model are as follows: evidence serobila can only be connected with chain head, and an evidence serobila can produce multiple chains
Head, a chain head can only correspond to an evidence serobila；One chain head can be connected with multiple tie-points, tie-point can with it is more
A chain head connection, a true node can be contacted with multiple tie-points.
3. evident information typing according to claim 1 and dividing method, it is characterised in that will have in step (1)
The original evidence text segmentation of multiple paragraphs is at several independent evident informations, the criterion terminated for an evidence description
Include: to encounter newline, encounter fullstop, and there is specific Chinese character serial number or the serial number of Arabic numerals to indicate as each
Demonstrate,prove the segmentation boundary etc. between text description.
4. evidence text elements according to claim 1 extract mode, it is characterized in that according to 4W1H key in step (2)
Elements recognition strategy, the keyword set for obtaining each evidence serobila and each fact (are calculated from each evidence text
Chain head short text and the tie-point short text being calculated from true text), it specifically includes:
Step (2.1) segments true and evident information, and anolytic sentence dependency structure relationship, and auxiliary uses regular expressions
Formula extracts key element What, i.e., the things being related in information；
Step (2.2) extracts key element When using regular expression from true and evident information, i.e., is related in information
Time；
Step (2.3) segments true and evident information, and analyzes part of speech, phrase structure relationship, extracts key element
Where, i.e., the place being related in information；
Step (2.4) segments true and evident information, and analyzes part of speech, extracts key element Who, i.e., involved in information
The party arrived；
Step (2.5) extracts key element How much using regular expression from true and evident information, i.e., relates in information
And the quantity arrived, it mainly include the amount of money and weight etc..
5. patterned modeling tool according to claim 1, three kinds of chain of evidence relational graph are drawn in modeling tool
Connection relationship, calculated result of three connection relationships of chain of evidence relational graph respectively by step (2), step (4) and step (5) are true
It is vertical, complete graphical tool is provided to support model definition and the rule in right 2.
6. the chain of evidence relational graph according to claim 1, drawn with chain of evidence relational graph modeling tool, in addition to visualized graphs
Outside storage form, there are also XML storage format, to support the chain of evidence relational model of the case to be loaded into different terminals computer and
Editor, conveniently handles a case.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811272642.1A CN109360127A (en) | 2018-10-24 | 2018-10-24 | A kind of chain of evidence relational graph modeling method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811272642.1A CN109360127A (en) | 2018-10-24 | 2018-10-24 | A kind of chain of evidence relational graph modeling method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109360127A true CN109360127A (en) | 2019-02-19 |
Family
ID=65347360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811272642.1A Pending CN109360127A (en) | 2018-10-24 | 2018-10-24 | A kind of chain of evidence relational graph modeling method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109360127A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110162607A (en) * | 2019-02-20 | 2019-08-23 | 北京捷风数据技术有限公司 | A kind of government organization document information retroactive method and device based on convolutional neural networks |
| CN110288495A (en) * | 2019-06-25 | 2019-09-27 | 北京市律典通科技有限公司 | Case statute of limitation intelligence checking method and device |
| CN110674840A (en) * | 2019-08-22 | 2020-01-10 | 中国司法大数据研究院有限公司 | Multi-party evidence association model construction method based on Bayesian network and evidence chain extraction method and device |
| CN110851591A (en) * | 2019-09-17 | 2020-02-28 | 河北省讯飞人工智能研究院 | Judgment document quality evaluation method, device, equipment and storage medium |
| CN115061982A (en) * | 2022-08-15 | 2022-09-16 | 四川科瑞软件有限责任公司 | Case-customization-based relational graph construction method, system, terminal and medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106970957A (en) * | 2017-03-17 | 2017-07-21 | 福州大学 | A kind of digital evidence chain overall analysis system and method |
| CN107632968A (en) * | 2017-05-22 | 2018-01-26 | 南京大学 | A kind of construction method of chain of evidence relational model towards judgement document |
| US20180082183A1 (en) * | 2011-02-22 | 2018-03-22 | Thomson Reuters Global Resources | Machine learning-based relationship association and related discovery and search engines |
| CN108595548A (en) * | 2018-04-09 | 2018-09-28 | 南京网感至察信息科技有限公司 | A kind of case judge's prediction of result method based on Markov Logic Network |
-
2018
- 2018-10-24 CN CN201811272642.1A patent/CN109360127A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180082183A1 (en) * | 2011-02-22 | 2018-03-22 | Thomson Reuters Global Resources | Machine learning-based relationship association and related discovery and search engines |
| CN106970957A (en) * | 2017-03-17 | 2017-07-21 | 福州大学 | A kind of digital evidence chain overall analysis system and method |
| CN107632968A (en) * | 2017-05-22 | 2018-01-26 | 南京大学 | A kind of construction method of chain of evidence relational model towards judgement document |
| CN108595548A (en) * | 2018-04-09 | 2018-09-28 | 南京网感至察信息科技有限公司 | A kind of case judge's prediction of result method based on Markov Logic Network |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110162607A (en) * | 2019-02-20 | 2019-08-23 | 北京捷风数据技术有限公司 | A kind of government organization document information retroactive method and device based on convolutional neural networks |
| CN110162607B (en) * | 2019-02-20 | 2021-08-31 | 北京捷风数据技术有限公司 | Government organization official document information tracing method and device based on convolutional neural network |
| CN110288495A (en) * | 2019-06-25 | 2019-09-27 | 北京市律典通科技有限公司 | Case statute of limitation intelligence checking method and device |
| CN110674840A (en) * | 2019-08-22 | 2020-01-10 | 中国司法大数据研究院有限公司 | Multi-party evidence association model construction method based on Bayesian network and evidence chain extraction method and device |
| CN110674840B (en) * | 2019-08-22 | 2022-03-25 | 中国司法大数据研究院有限公司 | Multi-party evidence association model construction method and evidence chain extraction method and device |
| CN110851591A (en) * | 2019-09-17 | 2020-02-28 | 河北省讯飞人工智能研究院 | Judgment document quality evaluation method, device, equipment and storage medium |
| CN115061982A (en) * | 2022-08-15 | 2022-09-16 | 四川科瑞软件有限责任公司 | Case-customization-based relational graph construction method, system, terminal and medium |
| CN115061982B (en) * | 2022-08-15 | 2022-10-25 | 四川科瑞软件有限责任公司 | Case-customization-based relational graph construction method, system, terminal and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109360127A (en) | A kind of chain of evidence relational graph modeling method | |
| Gardent et al. | Creating training corpora for nlg micro-planning | |
| Zhao et al. | Facilitating discourse analysis with interactive visualization | |
| Van Ham et al. | Mapping text with phrase nets | |
| CN108763483A (en) | A kind of Text Information Extraction method towards judgement document | |
| CN107632968A (en) | A kind of construction method of chain of evidence relational model towards judgement document | |
| Sun et al. | Text visualization for construction document information management | |
| CN111145052A (en) | Structured analysis method and system of judicial documents | |
| CN108564300B (en) | Efficient DWG drawing online collaborative verification method based on process engine | |
| Epure et al. | Automatic process model discovery from textual methodologies | |
| Li et al. | On vision features in multimodal machine translation | |
| KR20110009205A (en) | Systems and methods for natural language communication with a computer | |
| CN101661493A (en) | Interactive electronic technical manual browsing system | |
| CN115080751A (en) | Medical standard term management system and method based on general model | |
| CN116070599A (en) | Intelligent question bank generation and auxiliary management system | |
| Sevastjanova et al. | Questioncomb: A gamification approach for the visual explanation of linguistic phenomena through interactive labeling | |
| John et al. | A visual approach for the comparative analysis of character networks in narrative texts | |
| Seljan | Quality Assurance (QA) of Terminology in a Translation Quality Management System (QMS) in the business environment | |
| Pinquié et al. | Survey of Model-Based Design Reviews: Practices & Challenges? | |
| Pinheiro et al. | ChartText: Linking Text with Charts in Documents | |
| CN115204123B (en) | Collaborative editing document analysis method, analysis device, and storage medium | |
| Reggio et al. | DUSM: A method for requirements specification and refinement based on disciplined use cases and screen mockups | |
| CN108364168A (en) | A kind of administrative penalty data extraction system and method based on excel templates | |
| Ivanchyshyn et al. | The Film Script Generation Analysis Based on the Fiction Book Text Using Machine Learning | |
| Bethard | Finding event, temporal and causal structure in text: A machine learning approach |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190219 |
|
| WD01 | Invention patent application deemed withdrawn after publication |