KR101664454B1 - Apparatus and method for bulding law ontology - Google Patents

Apparatus and method for bulding law ontology Download PDF

Info

Publication number
KR101664454B1
KR101664454B1 KR1020150088868A KR20150088868A KR101664454B1 KR 101664454 B1 KR101664454 B1 KR 101664454B1 KR 1020150088868 A KR1020150088868 A KR 1020150088868A KR 20150088868 A KR20150088868 A KR 20150088868A KR 101664454 B1 KR101664454 B1 KR 101664454B1
Authority
KR
South Korea
Prior art keywords
legal
ontology
pattern
triple
sentence
Prior art date
Application number
KR1020150088868A
Other languages
Korean (ko)
Inventor
김명호
조대웅
Original Assignee
숭실대학교산학협력단
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 숭실대학교산학협력단 filed Critical 숭실대학교산학협력단
Priority to KR1020150088868A priority Critical patent/KR101664454B1/en
Application granted granted Critical
Publication of KR101664454B1 publication Critical patent/KR101664454B1/en

Links

Images

Classifications

    • G06F17/30734—
    • G06F17/2705—
    • 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)
  • Tourism & Hospitality (AREA)
  • Engineering & Computer Science (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Technology Law (AREA)
  • Primary Health Care (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Machine Translation (AREA)

Abstract

Disclosed are an apparatus and a method for building law ontology. The method for building law ontology builds ontology based on a law layer by extracting an item type included in the law, builds the ontology based on a law sentence pattern by converting the law sentence into triple by using previously defined law sentence pattern information and triple conversion regulation information, and builds the law ontology by mapping the ontology based on the law layer on the ontology based on the law sentence pattern.

Description

[0001] APPARATUS AND METHOD FOR BULDING LAW ONTOLOGY [0002]

The present invention relates to an apparatus and method for constructing a statutory ontology, and more particularly, to a statutory ontology constructing apparatus and method for defining a semantic relation between concepts included in a statute.

Law is universal knowledge and norm that human beings live. The law is the foundation of the nation in the rule of law, and has the characteristics of change, change, and change as time goes by. Everyone should know the law well in order to comply with norms and to ensure safety from crime.

In recent years, the government has provided web-based legal search services so that all citizens can read the legal texts. However, there is a problem that the law that should be approached to universal knowledge to all citizens is still difficult to use terms and approaches of legal knowledge for many people to use. It is not easy for ordinary people to understand legal knowledge, as well as legal terms, as well as misunderstanding the various relationships of the statute.

Therefore, it is necessary to construct a statutory ontology that defines the semantic relation between the concepts included in the statute structurally and systematically so that the general people can search the law more easily.

Japanese Patent Application Laid-Open No. 10-2012-0021011 Japanese Patent Application Laid-Open No. 10-2005-0051864 Japanese Patent Application Laid-Open No. 10-2014-0139909

One aspect of the present invention is to analyze and summarize the structural features of laws and regulations, to sort and sort sentence pattern patterns of statutory sentences, and to perform ontology mapping based on them to define the semantic relations between the concepts included in the statute An apparatus and method for constructing an ontology are provided.

According to one aspect of the present invention, a statutory ontology establishment method receives a statute, analyzes the statute, extracts at least one item type from the statute, constructs an ontology based on a statute hierarchy, Extracting main information using predefined statutory sentence pattern information, applying the extracted main information to predefined triple conversion information to generate a triple to construct an ontology based on a legal sentence pattern, And generates a statutory ontology by mapping the ontology and the legal sentence pattern based ontology.

In order to extract at least one item type from the statute, it is necessary to analyze at least one item among the items, sections, clauses, clauses, clauses, You can extract the item type for.

The ontology based on the statutory hierarchy is constructed by dividing the above-mentioned statute into a body composed of the head composed of the item type and a title or a sentence linked to the item type when the item type is extracted from the statute, Based ontology can be constructed.

The construction of the statutory hierarchy based ontology with the head and the body is performed by adding an item type included in the head to a node according to a vertical relationship between items included in the predefined statutory hierarchical type information to generate an item tree , It is possible to construct the statutory hierarchy based ontology by mapping bodies corresponding to the nodes.

The body may be stored as a title for the item type if the title exists for the item type, and may be stored as a content for the item type if the title for the item type does not exist.

The extracting of main information using the statutory sentence pattern information predefined in the legal sentences included in the statute compares the title of the item type included in the body with the predefined regulation class information, It is possible to check whether or not at least one of the predefined regulatory classes matches with each other, and extract important information according to the result.

The predefined regulatory class may include objectives, definitions, plans, committees, penalties, penalties, and enforcement date classes.

The extracting of the key information may include extracting the main information from the main body of the body if at least one of the title of the item type included in the body and the predefined regulated class match the legal sentence included in the body, Comparing the detected sentence pattern with a legal sentence pattern included in the body, comparing the detected sentence pattern with a legal sentence pattern included in the body, comparing the detected sentence pattern with a legal sentence pattern included in the body, The inconsistent portion can be extracted as the main information.

The generating of the triple may include: detecting triple conversion rules corresponding to legal statements included in the body by comparing the detected sentence pattern and the sentence pattern with triple conversion information of a predefined triple conversion rule, And generate the triple using the detected triple conversion rule.

The triple conversion information in which the triple conversion rule is predefined for each sentence pattern may store a descriptor used for triple conversion according to the sentence pattern.

Generating the triple using the extracted main information and the detected triple conversion rule includes extracting a predicate used for triple conversion of a legal sentence included in the body from a triple conversion rule corresponding to a legal sentence included in the body, And the triple can be generated by setting the extracted main information as a given or object and setting the detected predicate as a predicate.

The apparatus for constructing a statutory ontology according to an aspect of the present invention includes a TBox construction unit for analyzing a statute, extracting at least one item type from the statute, and constructing an ontology based on a statute hierarchy, The ontology based on the legal statement pattern is constructed by extracting the main information by using the statutory sentence pattern information, generating the triple by applying the extracted main information to the predefined triple conversion information, And an ABOX constructing unit that generates a statutory ontology by mapping an ontology based on a legal sentence pattern.

According to an aspect of the present invention, the statutory ontology that defines the semantic relation between the concepts included in the statute can be constructed to provide more meaningful legal information through the keyword-based search method.

1 is a block diagram of a statutory ontology building apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram for explaining the hierarchical type of statute.
FIG. 3 is a diagram illustrating an operation method of the TBox construction unit shown in FIG. 1. FIG.
4 is a diagram showing a legal domain ontology graph.
5 is a diagram showing an example of converting a legal sentence into a triple.
FIG. 6 is a flowchart illustrating a statutory ontology building method according to an embodiment of the present invention.

The following detailed description of the invention refers to the accompanying drawings, which illustrate, by way of illustration, specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that the various embodiments of the present invention are different, but need not be mutually exclusive. For example, certain features, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in connection with an embodiment. It is also to be understood that the position or arrangement of the individual components within each disclosed embodiment may be varied without departing from the spirit and scope of the invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is to be limited only by the appended claims, along with the full scope of equivalents to which such claims are entitled, if properly explained. In the drawings, like reference numerals refer to the same or similar functions throughout the several views.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the drawings.

The apparatus for building a statutory ontology according to an embodiment of the present invention can construct TBox, which is a knowledge base of the schema type, between statutory structures, extracts key information based on sentence pattern rules, And ABox which maps the extracted main information can be constructed to construct a statutory ontology.

FIG. 1 is a block diagram of an apparatus for constructing a statutory ontology according to an embodiment of the present invention. FIG. 2 is a diagram for explaining a hierarchy type of a statute, and FIG. 3 is a diagram illustrating an operation method of the TBox constructing unit shown in FIG. FIG. 4 is a diagram showing a legal domain ontology graph, and FIG. 5 is a diagram showing an example of converting a legal sentence into a triple.

Referring to FIG. 1, the statutory ontology construction apparatus 1 according to an embodiment of the present invention may include a statutory knowledge base 100 and a control unit 200.

The statutory knowledge base 100 can store statutory information. At this time, the statute may be composed of natural language, that is, non-standardized data. Also, the statute stored in the statute knowledge base 100 may mean a statute described in Korean.

In addition, the statutory knowledge base 100 can store statutory ontology information constructed by the control unit 200.

The control unit 200 can construct the ontology by analyzing the statutory statically and semantically. For this, a TBox construction unit 210 and an ABox construction unit 220 may be included.

In general, laws and ordinances are a combination of laws and ordinances, and they are divided into the name of the law, the principal and the appendices, and the attached documents are classified into separate tables, separate and attached forms. In the Act, the main text of this rule has eight levels of hierarchy, as shown in Fig. 2: chapters, chapters, chapters, chapters, clauses, clauses, and chapters. At this time, in order to distinguish each layer, the type of the item and the corresponding title are linked together. Referring to FIG. 2, an item type corresponding to a group can be classified by changing only the layer type after the numeric value in the form of Part 1, Part 1, Part 1, Part 1, and Part 1. However, the terms belonging to the subordinate clause are characterized by the notation in the form of '①', 'hoo', and '①'. The heading does not include a title, but rather describes the content of the article. In addition, laws and regulations are described in the introduction.

The TBox construction unit 210 can construct the relationship between the legal hierarchical relationship and the upper item of the clause in the legal document and the other laws with the OWL DL level schema.

Specifically, the TBox construction unit 210 can scan a legal document line by line. The TBox construction unit 210 can scan a legal document line by line and extract items included in a legal document. At this time, the TBox construction unit 210 can extract an item from the legal document according to the item extraction target type shown in [Table 1].

Figure 112015060585453-pat00001

The TBox construction unit 210 may scan the legal document line by line and extract an item corresponding to the item extraction target type shown in [Table 1]. The TBox construction unit 210 can structure the extracted items according to a hierarchical structure.

On the other hand, in the legal documents, sections, chapters, sections, chapters, chapters, sections, chapters, chapters, and chapters are not shown in the legal documents and most of them are described as chapters, clauses, clauses, It is common. Hereinafter, a description will be given of a method of structuring an item by, for example, the TBox construction unit 210 with the legal document described in the chapter, clause, clause,

Referring to FIG. 3, the contents shown on the left are schematically shown in the actual legal documents, and the data structure of the four-column arrangement on the upper right is an arrangement for distinguishing the levels of the items. The item tree on the lower right side shows the state when the processing of the item extracting operation is finished. First, only the root is present, and the level of the root May be zero.

When a legal document shown in FIG. 3 is scanned, a legal document may be divided into the following sections according to the items to be extracted in [Table 1]: 'chapter 1, section 1, section 1, section 2, Article 5, ①, ②, ③ can be read. At this time, since the TBox construction unit 210 is read first in the legal document when the 'chapter 1' is read, it recognizes that the 'chapter 1' corresponds to the highest priority level in the legal document, You can store the expression pattern 'chapter' in the first column of the item type precedence array. At this time, the pattern 'chapter' is stored in the first column of the item type priority array, which means it corresponds to level 1. Accordingly, the TBox construction unit 210 can structure the legal document by adding a node corresponding to the 'first chapter' to the level 1 of the item tree. At this time, the TBox construction unit 210 may add 'first chapter' as a child of the level 0, that is, 'root'.

In addition, when the legal document is scanned, 'Article 1' can be read after 'Chapter 1'. Since the 'first article' is read after the 'first article' in the legal document, the TBox construction section 210 generates the 'd' item of the 'first article' as the second item of the item type priority arrangement Lt; / RTI > At this time, the pattern 'd' is stored in the second column of the item type priority array, which means that it corresponds to level 2. Accordingly, the TBox construction unit 210 can structure a legal document by adding a node corresponding to the 'first set' to the level 2 of the item tree. At this time, the TBox construction unit 210 may add 'first set' as a child of the last added level 1, i.e., 'first chapter'.

'①' can be read next to 'Article 1'. The TBox construction unit 210 stores the regular expression pattern '[1 - 9]' in the third column of the item type priority order because '1' is read after '1' in the legal document . At this time, the pattern '[① - ⑨]' means that it corresponds to level 3 because it is stored in the third column of the item type priority array. Accordingly, the TBox construction unit 210 can structure a legal document by adding a node corresponding to '1' to the level 3 of the item tree. At this time, the TBox construction unit 210 may add '1' as a child of the last added level 2, that is, 'first set'.

Also, after '①' is read in legal documents, '②' can be read. '②', which is read next to '①', can be recognized as level type 3 as an item type corresponding to the regular expression pattern '[① - ⑨]' like '①'. Accordingly, the TBox construction unit 210 can add '2' as a child of level 2, that is, 'first set'. In this case, since '①' and '②' correspond to the same level of the same item type, '②' in the item tree as shown in FIG. 3 can be arranged on the same line as '①'.

In the legal documents, '②' followed by 'a.' Can be read. The TBox construction unit 210 can store the regular expression pattern 'a' in the fourth column of the item type priority arrangement because 'a' is read next to 'a' in the legal document. have. At this time, the pattern '[ha ha]' may mean that it corresponds to level 4 because it is stored in the fourth column of the item type priority array. Accordingly, the TBox construction unit 210 can structure a legal document by adding a node corresponding to '.' To level 4 of the item tree. At this time, the TBox construction unit 210 may add 'a' as a child of the last added level 3, that is, '2'. Also, in the legal document, 'I' and 'Da', which are read next to 'a', correspond to level 4, which corresponds to the regular expression pattern '[ha ha]', You can add 'I' and 'Da' to the added level 3, that is, the children of '②'.

If the 'chapter 2' is read during the legal document scan shown in FIG. 3, the 'chapter d', which is the regular expression pattern of the 'chapter 2' through the item type priority arrangement generated through the previous process, It can be recognized as such. Accordingly, the TBox construction unit 210 can add a node for the 'second chapter' to the tree corresponding to the level 1. At this time, 'Chapter 2' can be added as a child of 'root' level 0. In addition, the TBox construction unit 210 may extract the other items to be read next to the 'second chapter' according to the type in the same manner as described above, and store the extracted items in a tree structure. At this time, the TBox construction unit 210 can set the item part as the head part in the line of the selected type in the item type, set the body part as the part excluding the item, and store it as the title of the corresponding item. For example, for "Article 2 (extent of survivor)", "Article 2" corresponds to item type, so "Article 2" can be set as head part, and "instance 2" You can set the title part to the body part and save it as "(survivor's range)". On the other hand, since the type corresponding to an item, a head, or a neck is a type without a title, the part starting with a body becomes a form of a legal clause, so it can be stored as a legal form without being classified as a title.

The TBox construction unit 210 scans the legislation line by line and extracts the items corresponding to the type of the line, chapter, section, line, line, line, arc, and tree as described above, Tree, it is possible to construct a TBox based on statute hierarchy.

On the other hand, the legal domain ontology can be composed of two classes, namely, one class and a regular class, as shown in FIG. A class is an item at the top level of the legal document system, and a subclass may include a class of items such as chapter, clause, clause, clause, clause, clause, and clause level that conform to the legal document system. Each class can be mapped to an item information instance associated with the item. In addition, the subclasses related to the text can be matched in relation to the subclasses and the object properties. At this time, the matching of the class and the rule class is because the actual legal document shows the form of the legal sentence related to the actual rule starting from the sentence, the sentence element is matched with the instance of the prescribed class according to the item, It can be. Such a relationship mapping is intended to enable the extraction of related concepts by inference in the execution of a legal sentence query. A detailed description of the relationship mapping will be described later through the ABox constructing unit 220. [

In addition, a rule class, which is one of the upper classes of the legal domain ontology, is a superclass of rules that can be extracted from legal documents, and subclasses can include classes such as purpose, definition, committee, plan, penalty, . On the other hand, the legal documents include purpose, definition, committee, plan, penalties, fines, and other regulatory classes other than the enforcement date. However, the regulatory class according to an embodiment of the present invention includes purposes, definitions, committees, Penalty, penalty, and enforcement date. An instance of each class can be constructed from an ontology by extracting it as a predicate, object form, given a triple that meets the purpose of the regulation from the legal text. A detailed description thereof will be given later through the ABox constructing unit 220. [

The ABox constructing unit 220 can extract the main information from the statute based on the legal sentence pattern, and establish the relationship between the extracted main information and the instance axiom of TBox to construct the on-law ontology.

Specifically, the ABox construction unit 220 can extract legal statements from legal documents. The ABox construction unit 220 can analyze the extracted legal sentence and detect which sentence pattern corresponds to the predetermined sentence pattern. The ABox construction unit 220 can detect which rule is included in the legal statement. The ABox construction unit 220 can detect the pattern number corresponding to the legal statement. At this time, the sentence pattern, regulation, and pattern number of the corresponding legal sentence can be detected using the following [Table 2].

Pattern number Rule Sentence pattern Pattern 1 Purpose The law aims at NLo Pattern 2 Definition "NLs" means NLo Pattern 3 "NLs" means NLo Pattern 4 &Quot; NLs " Pattern 5 &Quot; NLs " Pattern 6 NLo {lower layer} Pattern 7 NLo {NLo1, NLo2, NLo3, ...} Pattern 8 Plan Secretary NLo Pattern 9 NLo Chang Pattern 10 NLo years Pattern 11 (Hereinafter referred to as the "NLo Plan"), Pattern 12 NLo {lower layer} Pattern 13 Determine by NLo Pattern 14 Notify NLo Pattern 15 Committee NLo Committee Pattern 16 The chairman becomes NLo Pattern 17 Chairman NLo Pattern 18 NLo members or less Pattern 19 Type 6 NL -> NLo {lower layer} Pattern 20 Type 6 NL Person in each of the following -> NLo {lower layer} Pattern 21 Type 6 NL -> NLo {lower layer} Pattern 22 Type 6 Review the following issues -> NLo {lower layer} Pattern 23 Determine by NLo Pattern 24 The term of the committee is NLo Pattern 25 Penalty Type 6 NLos people Pattern 26 In the case of Type 6 NLos Pattern 27 Type 6 NLos -> NLo {lower layer} Pattern 28 For Type 6 NLos -> NLo {lower layer} Pattern 29 NLo prison sentence Pattern 30 Prison sentence of less than NLo Pattern 31 Fine over NLo Pattern 32 Fine below NLo Pattern 33 NL detention or fine Pattern 34 NL detention Pattern 35 NL fine Pattern 36 Fine (fine) For NLos people Pattern 37 In the case of NLos Pattern 38 For NLos, Pattern 39 For Type 6 NLos people -> NLo {lower layer} Pattern 40 For Type 6 NLos -> NLo {lower layer} Pattern 41 For Type 6 NLos, the -> NLo {lower layer} Pattern 42 Penalty of NLo or less Pattern 43 I can not dispose of my fines Pattern 44 Enforcement date NLo months after fear Pattern 45 The day of promulgation Pattern 46 Legal Type Type 1 NLo Pattern 47 Type 2 NLo Pattern 48 Type 3 NLo Pattern 49 Type 4 NLo Pattern 50 Type 5 (NLo) Pattern 51 Type 6 NLo Pattern 52 Type 6 NLo: Pattern 53 Type 7 NLo Pattern 54 Type 7 NLo: Pattern 55 Type 8 NLo Pattern 56 Type 8 NLo:

The legal sentence pattern shown in [Table 2] may be a pattern that summarizes the extractable information of the common sentence that appears by the rule and the law level. In Table 2, NL {s, o, os} is abbreviation of Natural Language. Subscript s is a Subject, o is an Object, and os is an ObjectSubject will be. On the other hand, in [Table 2], patterns 1 through 45 correspond to the legal sentence patterns corresponding to the prescribed class, and patterns 46 through 56 may compose legal sentence patterns corresponding to the legal hierarchy class.

5, the " " public data " refers to data that is generated or acquired by a public institution such as a database or an electronic file for the purpose specified by laws and ordinances, Data or information "is read, a sentence pattern corresponding to the corresponding legal sentence can be detected by comparing the sentence pattern indicated in [Table 2] with the corresponding legal sentence. At this time, "" public data "refers to data or information processed by optical or electronic means, such as databases, electronic files, etc., which are created, acquired and managed by public organizations for purposes specified by laws and ordinances. The pattern may be detected as " NLs " NLo ". Further, the ABox construction unit 220 detects a sentence pattern corresponding to the legal sentence, and detects that the corresponding sentence corresponds to the " definition " rule and corresponds to pattern 2 using the detected sentence pattern .

In addition, the ABox construction unit 220 may detect a sentence pattern corresponding to a legal sentence, and then extract key information from the sentence pattern using the detected sentence pattern. At this time, the ABox construction unit 220 compares the sentence pattern of the legal sentence with the corresponding sentence pattern, and extracts the main information of the word or the sentence except the part that matches the sentence pattern of the corresponding sentence in the legal sentence. For example, the term "public data" refers to data or information processed in an optical or electronic way that is created, acquired and managed by a public entity such as a database or an electronic file for purposes specified by laws and ordinances. The corresponding sentence pattern is "NLs" and "NLo". Words or phrases excluding the parts that match the pattern when they are compared are "public data" and "databases, electronic files, Quot; refers to data or information processed in an optical or electronic manner that is generated or acquired and managed for the purposes of the present invention. &Quot; The ABox construction unit 220 refers to data or information processed in an optical (optical) or electronic manner that is generated or acquired and managed by a public entity such as a database, an electronic file, etc. for the purpose specified by laws and ordinances "And extracts" public data "as NLs according to the sentence pattern of the relevant legal sentence, and" extracts public data "to NLs for the purpose of the public institution such as database, Or data or information processed in an optical or electronic manner that is acquired and managed. "Can be extracted as NLo.

The ABox construction unit 220 can convert the legal sentence into a triple using the pattern information of the legal sentence and the main information extracted from the legal sentence. At this time, the ABox construction unit 220 can convert the legal sentence into triple using the triple conversion rule shown in the following [Table 3]. The triple transfer rules shown in [Table 3] are rules in which predefined predicates are mapped according to patterns. At this time, by mapping predefined predicates according to the regulatory information, it is possible to semantically confirm what information the corresponding triple represents through the name of the predicate. Such a method may be useful in the future when a legal information retrieval system wants to find information based on a predicate.

Rule number Pattern number subject terminology direct object Rule 1 Pattern 1 Act klaw: isPurposeTo NLo Rule 2 Patterns 2, 3, 4, 5 NLs rdfs: isDefinedBy NLo Rule 3 Pattern 6, 7 kllaw: kindOf Rule 4 Pattern 8, 9 Plan klaw: minister NLo Rule 5 Pattern 10 klaw: year Rule 6 Pattern 11 klaw: planType Rule 7 Pattern 12 klaw: include Rule 8 Pattern 13 klaw: pIndicator Rule 9 Pattern 14 klaw: notice Rule 10 Pattern 15 Committee klaw: committee NLo Rule 11 Pattern 16, 17 klaw: chairPeson Rule 12 Pattern 18 klaw: committeeNumber Rule 13 Pattern 19, 20 klaw: eligibilityCommittee Rule 14 Pattern 21, 22 klaw: deliberation Rule 15 Pattern 23 klaw: cIndicator Rule 16 Pattern 24 klaw: termCommittee Rule 17 Patterns 25, 26, 27, 28 penalty klaw: penaltyRelevantIems NLos Rule 18 Pattern 29 NLos klaw: prisonLaborAbove NLo Rule 19 Pattern 30 klaw: prisonLaborBelow Rule 20 Pattern 31 klaw: penaltyAbove Rule 21 Pattern 32 klaw: penaltyBelow Rule 22 Pattern 33 klaw: penaltyRegulations Detention or fine Rule 23 Pattern 34 detention Rule 24 Pattern 35 mulct Rule 25 Pattern 36, 37, 38, 39, 40, 41 Penalty klaw: fineRelevantItems NLos Rule 26 Pattern 42 NLos klaw: fine NLo Yuan Rule 27 Pattern 43 No disposal Rule 28 Pattern 44 enforcement klaw: proclamation NLo months Rule 29 Pattern 45 The day of promulgation Rule 30 Pattern 46 Type 1 Klaw: rule NLo Rule 31 Pattern 47 Type 2 Rule 32 Pattern 48 Type 3 Rule 33 Pattern 49 Type 4 Rule 34 Pattern 50 Type 5 Rule 35 Pattern 51 Type 6 klaw: siProvision Rule 36 Pattern 52 klaw: siKeyword Rule 37 Pattern 53 Type 7 klaw: seProvision Rule 38 Pattern 54 klaw: seKeyword Rule 39 Pattern 55 Type 8 klaw: eProvision Rule 40 Pattern 56 klaw: eKeyword

The ABox construction unit 220 can detect the predicate corresponding to the corresponding legal sentence pattern through [Table 3]. The ABox construction unit 220 can convert the legal sentence into triple with the main information extracted through [Table 2] and the predicate extracted through [Table 3]. For example, the phrase "public data" refers to data or information processed in an optical or electronic manner that is created, acquired and managed by a public entity, such as a database, an electronic file, Pattern number 2, so that it can be confirmed that it corresponds to the triple conversion rule 2 through [Table 3]. Accordingly, a predefined predicate corresponding to the legal sentence can be detected as " isDefinedBy ". As shown in FIG. 5, "subject data" includes "public data", predicate "isDefinedBy" ) May be converted into a triple consisting of "data or information processed in an optical or electronic manner, such as databases, electronic files, etc., created or acquired by a public entity for purposes specified by laws and ordinances".

The ABox construction unit 220 can generate a statutory ontology by mapping the transformed triple and the corresponding legal domain ontology.

Specifically, the ABox construction unit 220 can detect a class to which a legal statement corresponding to the converted triple belongs. The ABox construction unit 220 can map the detected class and triple. At this time, the statutory ontology can be generated by mapping the rule class corresponding to the triple and the class class by mapping the class class and the triple.

Hereinafter, a statutory ontology building method according to an embodiment of the present invention will be described with reference to FIG.

First, a legal sentence of a legal document is inputted (310) from a statutory knowledge base (100) and an inputted legal sentence is analyzed line by line (315).

When analyzing a legal sentence on a line-by-line basis, the head of the line is compared (320) to see if the head of the line matches the item type within the law.

At this time, the legal item type can mean the item type shown in [Table 1].

If the head of the line is verified (320) to match the legal item type, it is recognized as a legal document and the legal document is separated into head and body (325).

At this time, the head can refer to items in the legal document that match the legal item type, and the body can include the title or contents of the article linked to the item.

The item tree for the legal document is created using the separated head and body (330).

In this case, since the legal document is described as a long form, the order of the items read in the legal document represents the hierarchical structure of the items, so that the item tree can be generated by arranging the item types according to the order of the items read in the legal document have. On the other hand, each item can generate an item tree as an item type, and each item constitutes a part for starting the body as the item's title if the item's title exists, and if the item's title does not exist The subject, and the subject), it is possible to construct an item tree by constructing a part starting the body with the contents of the article, and mapping the body corresponding to each item.

(335) whether the beginning of the body of the class corresponding to the article indicating the actual legal content matches the predetermined class of the prescribed rule.

In this case, the predefined rule class may be a purpose, a definition, a plan, a committee, a penalty, a penalty, an effective date class which are common in the legal document and confirming whether the beginning of the body corresponds to a predefined rule class May be to confirm whether the title of the document matches the purpose, definition, plan, committee, penalty, penalty, or enforcement date.

If it is confirmed that the body start of the jog class matches the predetermined rule class (335), a triple is generated according to a predefined legal sentence pattern (340).

In this case, to generate a triple according to a legal sentence pattern, a legal sentence included in the body is referred to a corresponding sentence in a sentence of [Table 2] in which a legal sentence pattern is defined according to purpose, definition, plan, committee, penalty, A corresponding sentence pattern is detected, and a triple conversion rule corresponding to the sentence pattern detected through [Table 3] in which the triple conversion rule is defined according to the sentence pattern is detected, and the corresponding legal sentence is converted into a triple, Can be generated.

Also, if the beginning of the body of the class is identified (335) to be inconsistent with a predetermined class, ie the class does not belong to any one of the purpose, definition, plan, committee, penalty, penalty, or enforcement date If it is confirmed, it is checked whether the item type corresponding to the item, the call, and the neck is read from the class body (345)

In this case, the item types corresponding to the terms, ①, ②, ③, ... , 1., 2., 3., ... , ABC., … Lt; / RTI >

When the item type corresponding to the item, the call, and the item is read (345) in the body of the class, a triple is generated according to a predefined legal hierarchy pattern (350).

To generate a triple according to a pre-defined legal hierarchy pattern, a sentence pattern corresponding to the body of the item is detected through [Table 2], and a triple conversion rule is defined according to the detected sentence pattern [ ] To detect the triple conversion rule corresponding to the detected sentence pattern, and convert the information included in the body of the corresponding class into triples.

In addition, if the item type corresponding to the item, the call, and the neck is not read in the body of the class, the part is tagged and parsed in the body part of the class through the morpheme analyzer to generate the triple (355).

After the triple generation (340, 350, 355), the legal domain ontology corresponding to the corresponding triple is mapped to generate the legal ontology (360).

Such a technique for constructing an ontology for a statute described in Korean can be implemented in an application or in the form of program instructions that can be executed through various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program commands, data files, data structures, and the like, alone or in combination.

The program instructions recorded on the computer-readable recording medium may be ones that are specially designed and configured for the present invention and are known and available to those skilled in the art of computer software.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like.

Examples of program instructions include machine language code such as those generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules for performing the processing according to the present invention, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it will be understood by those skilled in the art that various changes and modifications may be made therein without departing from the spirit and scope of the invention as defined in the appended claims. It will be possible.

100: Legal knowledge base
210: TBox building part
220: ABox building part

Claims (12)

The TBox builder receives the input of the decree,
Wherein the TBox construction unit analyzes the legislation on a line by line basis and extracts at least one item type from one of a piece, a chapter, a section, a section, a clause, an argument, a call, and a neck in accordance with a statutory hierarchy type information defined in advance in the Act, An ontology based on a statute hierarchy is constructed by dividing a statute into a body composed of a head composed of the item type and a title composed of the item type or the contents of the article,
The ABox construction unit extracts key information using pre-defined statutory sentence pattern information in a legal sentence included in the above statute, generates a triple by applying the extracted main information to predefined triple conversion information, And ontology of
Wherein the ABOb construction unit generates a statutory ontology by mapping the statutory hierarchy based ontology and the legal sentence pattern based ontology. delete delete The method according to claim 1,
The construction of the statutory hierarchy-based ontology with the head and the body,
Based on a hierarchical relationship between items included in the predefined statutory hierarchy type information, generates an item tree by adding an item type included in the head as a node, maps a body corresponding to each node, Constructing ontology Constructing ontology. The method according to claim 1,
The body
If the title for the item type exists, the body is stored as a title for the item type, and if the title for the item type does not exist, the body is stored as a content for the item type. 6. The method of claim 5,
In extracting the main information by using the statutory sentence pattern information predefined in the legal sentences included in the above-mentioned statute,
Comparing the part of the body with predefined regulatory class information to determine whether at least one of the predefined regulatory class and the title of the item type contained in the body matches with each other, How to build an ontology. The method according to claim 6,
The predefined regulatory class includes a purpose, a definition, a plan, a committee, a penalty, a penalty and an enforcement date class. The method according to claim 6,
In extracting the main information,
If at least one of the title of the item type included in the body matches with the predefined rule class, the pre-defined legal sentence pattern information is compared with the legal sentence pattern information according to the legal sentence included in the body and the rule, And comparing the detected sentence pattern with a legal sentence included in the body, and comparing the detected sentence pattern with the detected sentence pattern in the legal sentence included in the body to the main information How to build a statute ontology to extract. Claim 9 has been abandoned due to the setting registration fee. 9. The method of claim 8,
To generate the triple,
A triple conversion rule predefined by the triple conversion rule is detected for each sentence pattern and a sentence pattern to detect a triple conversion rule corresponding to a legal sentence included in the body, and the detected main information and the detected triple conversion rule To generate the triple using the statutory ontology. Claim 10 has been abandoned due to the setting registration fee. 10. The method of claim 9,
Wherein a triple conversion rule in which a triple conversion rule is predefined for each sentence pattern is stored in a predicate used for triple conversion for each sentence pattern. Claim 11 has been abandoned due to the set registration fee. 11. The method of claim 10,
Generating the triple using the extracted main information and the detected triple conversion rule,
Detecting a predicate used for triple conversion of a legal sentence included in the body from a triple conversion rule corresponding to a legal sentence included in the body, setting the extracted main information as a subject or an object term, To generate the triple. The system analyzes the statute in line and extracts at least one item type from among pieces, sections, clauses, clauses, clauses, clauses, and necks according to the statutory hierarchy type information defined in advance in the above statute, A TBox constructing unit for constructing an ontology based on a statutory hierarchy by separating a head composed of contents of the head and the contents linked to the item type; And
The main information is extracted by using the statutory sentence pattern information predefined in the legal sentences included in the above-mentioned statute, and the extracted main information is applied to the predefined triple conversion information to generate a triple to generate an ontology based on a legal sentence pattern And an ABOX constructing unit for generating a statutory ontology by mapping the ontology based on the statutory hierarchy and the ontology based on the legal sentence pattern.
KR1020150088868A 2015-06-23 2015-06-23 Apparatus and method for bulding law ontology KR101664454B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020150088868A KR101664454B1 (en) 2015-06-23 2015-06-23 Apparatus and method for bulding law ontology

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020150088868A KR101664454B1 (en) 2015-06-23 2015-06-23 Apparatus and method for bulding law ontology

Publications (1)

Publication Number Publication Date
KR101664454B1 true KR101664454B1 (en) 2016-10-10

Family

ID=57145705

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020150088868A KR101664454B1 (en) 2015-06-23 2015-06-23 Apparatus and method for bulding law ontology

Country Status (1)

Country Link
KR (1) KR101664454B1 (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102000546B1 (en) 2018-09-20 2019-07-17 한국건설기술연구원 Building regulation information search system with mapping and hyperlinking the commissioning relation between provisions of law and subordinate statute, and method for the same
KR102009649B1 (en) 2018-08-21 2019-08-13 한국건설기술연구원 Construction regulation legal information search system according to each classification plan of construction regulation, and method for the same
KR20220040697A (en) 2020-09-24 2022-03-31 이후록 System for establishment of relational network between provisions and multiviewer
KR20220091761A (en) 2020-12-24 2022-07-01 한국건설기술연구원 Computerized managing system for deriving building authorizing regulations and for providing checklist of each regulations, and method for the same

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20050051864A (en) 2003-11-28 2005-06-02 한국전자통신연구원 System and method for knowledge modeling using ontology
KR20120021011A (en) 2010-08-31 2012-03-08 대한민국(법무부장관) Intergrated law information service system and method
KR20140139909A (en) 2013-05-28 2014-12-08 숭실대학교산학협력단 Teminology ontology search service offering apparatus and the method thereof

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20050051864A (en) 2003-11-28 2005-06-02 한국전자통신연구원 System and method for knowledge modeling using ontology
KR20120021011A (en) 2010-08-31 2012-03-08 대한민국(법무부장관) Intergrated law information service system and method
KR20140139909A (en) 2013-05-28 2014-12-08 숭실대학교산학협력단 Teminology ontology search service offering apparatus and the method thereof

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
심준 외, 검색 키워드 확장을 이용한 온톨로지 자동 생성 시스템 개발, 한국산학기술학회논문지, Vol.10 No.6, pp.1220-1228, 2009. *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102009649B1 (en) 2018-08-21 2019-08-13 한국건설기술연구원 Construction regulation legal information search system according to each classification plan of construction regulation, and method for the same
KR102000546B1 (en) 2018-09-20 2019-07-17 한국건설기술연구원 Building regulation information search system with mapping and hyperlinking the commissioning relation between provisions of law and subordinate statute, and method for the same
KR20220040697A (en) 2020-09-24 2022-03-31 이후록 System for establishment of relational network between provisions and multiviewer
KR102551949B1 (en) * 2020-09-24 2023-07-06 이후록 System for establishment of relational network between provisions and multiviewer
KR20220091761A (en) 2020-12-24 2022-07-01 한국건설기술연구원 Computerized managing system for deriving building authorizing regulations and for providing checklist of each regulations, and method for the same

Similar Documents

Publication Publication Date Title
Sleiman et al. Trinity: on using trinary trees for unsupervised web data extraction
CN103229162B (en) Candidate answers logic synthesis is used to provide problem answers
Sannier et al. An automated framework for detection and resolution of cross references in legal texts
EP1665080A1 (en) Improved search engine
KR101664454B1 (en) Apparatus and method for bulding law ontology
Hazber et al. Integration mapping rules: Transforming relational database to semantic web ontology
de Boer et al. Linking the kingdom: enriched access to a historiographical text
Freitas et al. A Semantic Best-Effort Approach for Extracting Structured Discourse Graphs from Wikipedia.
Sun A natural language interface for querying graph databases
Tatu et al. Automatic extraction of actionable knowledge
Westphal et al. Countering language attrition with PanLex and the Web of Data
Zhekova et al. An Algorithm for Translation of a Natural Language Question into SQL Query.
KR101072147B1 (en) Method and system for transforming blog posting to Ontology-based information
Kumar et al. Keyword based search and its limitations in the patent document to secure the idea from its infringement
US10042928B1 (en) System and method for automated reasoning with and searching of documents
KR100504632B1 (en) Apparatus for extracting information desired by users from unstructured documents and method thereof
Aljalbout et al. A semantic model for historical manuscripts
Jain et al. Constructing a Knowledge Graph from Indian Legal Domain Corpus.
Fudholi et al. Code (common ontology development): A knowledge integration approach from multiple ontologies
Loza Mencía Segmentation of legal documents
Khalil et al. Challenges in information retrieval from unstructured arabic data
Shaburova Operational Experience in DB
Geetha et al. Extraction of key attributes from natural language requirements specification text
Vickers Ontology-based free-form query processing for the semantic web
Ibrahim et al. Plagiarism Detection Techniques for Arabic Script Languages: A Literature Review

Legal Events

Date Code Title Description
E701 Decision to grant or registration of patent right
GRNT Written decision to grant
FPAY Annual fee payment

Payment date: 20191001

Year of fee payment: 4