Document analysis support device, method and program

Abstract

To provide a document analysis support apparatus, a document analysis support method, and a document analysis support program, for correcting a rule appropriately.SOLUTION: A document analysis support apparatus D stores one or more combined words as co-occurrence rules, stores one or more first sentences as important sentences, stores one or more second sentences, which are different from the first sentences, as non-important sentences, determines whether the stored important sentences are important on the basis of the store co-occurrence rules, displays important sentences determined to be not important as a result of the determination on a display unit 3, receives input of multiple words as candidates of co-occurrence rule for the displayed important sentences, determines whether the stored non-important sentences are important on the basis of the received candidates of co-occurrence rule, and displays the received candidates of co-occurrence rule and a result determined to be not important in the determination on importance on the display unit 3.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a document analysis support device, a document analysis support method, and a document analysis support program that support the analysis of a document. [Background technology]

[0002] If multiple sentences that make up a document could be automatically analyzed, it would be convenient because it would reduce the time required to understand and check the content of the document, and it would standardize (objectiveize) the understanding and checking. Such technology for automatic document analysis is disclosed, for example, in Patent Document 1.

[0003] The document creation support device disclosed in Patent Document 1 has a memory unit in which rules that are conditions for extracting points of concern are stored, and a control unit that extracts points of concern from the titles of clauses in a document and sentences contained in the clauses based on the rules, determines whether the titles correspond to points of concern, and, if the determination shows that the titles correspond to points of concern, highlights the titles and sentences contained in the clauses on the screen regardless of whether the sentences contain points of concern. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-115858 Summary of the Invention [Problem to be solved by the invention]

[0005] According to paragraph

[0056] of Patent Document 1, the document creation support device disclosed in Patent Document 1 uses rules, which are conditions for extracting critical passages, to achieve the following effect: "For example, even if the clause title itself is a critical passage, or even if the clause content contains many individual critical passages, it is possible to highlight the clause title and all sentences included in the clause content. As a result, the burden on the contracting parties (creators) is reduced. Furthermore, since the critical passages can be edited, the user can freely select, as needed, between highlighting the clause title 102 and all sentences, or highlighting individual critical passages." However, since critical passages differ depending on the type of document, it is thought that the rules need to be modified depending on the type of document, but Patent Document 1 does not describe or suggest this rule modification.

[0006] The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a document analysis support device, a document analysis support method, and a document analysis support program that are capable of appropriately correcting rules. [Means for solving the problem]

[0007] As a result of various investigations, the present inventors have found that the above object can be achieved by the present invention described below. That is, a document analysis support device according to one embodiment of the present invention includes a display unit, a co-occurrence rule memory unit that stores one or more combined words as co-occurrence rules, an important sentence memory unit that stores one or more first sentences as important sentences, a non-important sentence memory unit that stores one or more second sentences different from the first sentences as non-important sentences, an importance judgment unit that determines whether an important sentence stored in the important sentence memory unit is important based on the co-occurrence rules stored in the co-occurrence rule memory unit, a display processing unit that displays important sentences that are determined to be unimportant by the importance judgment unit on the display unit, an input unit that accepts input of multiple words as candidate co-occurrence rules for the important sentences displayed on the display unit, and a non-importance judgment unit that judges whether a non-important sentence stored in the non-important sentence memory unit is important based on the candidate co-occurrence rules accepted by the input unit, and the display processing unit displays the candidate co-occurrence rules accepted by the input unit and the judgment result of the non-important sentences that are determined to be important on the display unit. Preferably, the document analysis support device described above further includes a second input unit that accepts an input of an instruction to add a new co-occurrence rule candidate, and a control unit that, when the second input unit accepts the instruction, adds the co-occurrence rule candidate as a new co-occurrence rule to the co-occurrence rule storage unit and stores it. Preferably, in the document analysis support device described above, the important sentence is a sentence that is included in a contract and has been determined to be important in advance, and the unimportant sentence is a sentence that is included in the contract and has been determined to be unimportant in advance. Preferably, in the document analysis support device described above, the important sentence is a sentence that is included in a specification and has been determined to be important in advance, and the unimportant sentence is a sentence that is included in the specification and has been determined to be unimportant in advance. Preferably, in the document analysis support device described above, the important sentence is a sentence that is included in an instruction manual and has been determined to be important in advance, and the unimportant sentence is a sentence that is included in the instruction manual and has been determined to be unimportant in advance.

[0008] When an important sentence is determined to be unimportant based on the co-occurrence rules stored in the co-occurrence rule storage unit, this document analysis support device displays the unimportant important sentence on the display unit, allowing the user to input a candidate co-occurrence rule for the important sentence determined to be unimportant while referring to the important sentence displayed on the display unit. The document analysis support device determines whether or not a non-important sentence stored in the non-important sentence storage unit is important based on the candidate co-occurrence rules, and displays the candidate co-occurrence rules and the determination result of the importance on the display unit. Therefore, the candidate co-occurrence rules can be adjusted while referring to the determination result so that a certain sentence is not determined to be important. Therefore, the document analysis support device can appropriately modify the co-occurrence rules.

[0009] The importance determination unit determines whether or not important sentences stored in the important sentence storage unit are important, and the unimportance determination unit determines whether or not unimportant sentences stored in the unimportance sentence storage unit are important. The importance determination unit and the unimportance determination unit only differ in the sentences they determine as important, but they use the same logic for determining whether a sentence is important, and both determine that a sentence is important if the co-occurrence rules match the sentence they determine as important.

[0010] In another aspect, in the document analysis support device, the determination result includes the number of unimportant sentences determined to be important by the unimportance determination unit.

[0011] Such a document analysis support device displays the number of unimportant sentences that have been determined to be important, and therefore, the co-occurrence rule candidates can be adjusted so that this number becomes zero.

[0012] In another aspect, in the above-mentioned document analysis support device, the display processing unit displays an operation button on the display unit for displaying non-important sentences determined to be important by the non-importance judgment unit on the display unit, and when the operation button is operated, the non-important sentences determined to be important by the non-importance judgment unit are displayed on the display unit.

[0013] Such a document analysis support device displays non-important sentences that have been determined to be important on a display unit, and can therefore efficiently adjust the candidates for co-occurrence rules while referring to the words contained in the non-important documents so that the non-important sentences are not determined to be important (so that the non-important sentences are not extracted).

[0014] In another aspect, the document analysis support device described above further includes an ontology storage unit that stores an ontology that represents the relationship between words, and the importance determination unit determines whether or not an important sentence stored in the important sentence storage unit is important based on the co-occurrence rules stored in the co-occurrence rule storage unit and the ontology stored in the ontology storage unit.

[0015] Such a document analysis support device determines whether something is important or not based on an ontology, and therefore can appropriately determine that a word contained in an important sentence is important even if it is a paraphrased word.

[0016] In another aspect, in the above-mentioned document analysis support device, the display processing unit displays on the display unit the relationships between multiple words as candidates for co-occurrence rules received by the input unit, based on the ontology stored in the ontology memory unit.

[0017] Such a document analysis support device displays words related (associated) with a plurality of words as candidates for co-occurrence rules, and therefore, it becomes possible to adjust the candidates for co-occurrence rules using the related words.

[0018] In another aspect, in the document analysis support device described above, the input unit accepts changes to multiple words as candidate co-occurrence rules that have already been accepted by the input unit, the unimportance determination unit re-determines whether or not a non-important sentence stored in the non-important sentence storage unit is important based on the candidate co-occurrence rules after the change accepted by the input unit, and the display processing unit displays the candidate co-occurrence rules after the change accepted by the input unit and the re-determination result that has been judged to be important by the unimportance determination unit on the display unit.

[0019] Such a document analysis support device changes the words of the co-occurrence rule candidates and re-evaluates using the changed co-occurrence rule candidates, so it can recognize the appropriateness of the changed co-occurrence rule candidates and efficiently adjust the co-occurrence rule candidates.

[0020] In another aspect, the document analysis support device further includes a co-occurrence rule modification unit that modifies the co-occurrence rules stored in the co-occurrence rule storage unit in accordance with related words related to the important sentences stored in the important sentence storage unit, and the importance determination unit determines whether the important sentences stored in the important sentence storage unit are important based on the co-occurrence rules modified by the co-occurrence rule modification unit. Preferably, in the document analysis support device, the related words are first words included in the title of a document consisting of the important sentences stored in the important sentence storage unit or second words related to the first words. Preferably, in the document analysis support device, the co-occurrence rule modification unit modifies the co-occurrence rules by deleting some of the words included in the co-occurrence rules in accordance with the related words. Preferably, in the document analysis support device, the co-occurrence rule modification unit modifies the co-occurrence rules by further combining at least one or more first and second words with the words included in the co-occurrence rules in accordance with the related words.

[0021] Such a document analysis support device modifies the co-occurrence rules in accordance with the related words related to the important sentences stored in the important sentence storage unit, and therefore can determine the importance more appropriately.

[0022] In another aspect, in the document analysis support device described above, the related words are further related to unimportant sentences stored in the unimportant sentence storage unit, the co-occurrence rule correction unit corrects the co-occurrence rule candidates received by the input unit in accordance with the related words, and the unimportance determination unit determines whether the unimportant sentences stored in the unimportant sentence storage unit are important based on the co-occurrence rule candidates corrected by the co-occurrence rule correction unit. Preferably, in the document analysis support device described above, the related words are a first word or a second word related to the first word included in a title of a document comprising an important sentence stored in the important sentence storage unit and an unimportant sentence stored in the unimportant sentence storage unit. Preferably, in the document analysis support device described above, the co-occurrence rule correction unit corrects the co-occurrence rule candidates by deleting some of the words included in the co-occurrence rule candidates in accordance with the related words. Preferably, in the document analysis support device described above, the co-occurrence rule correction unit corrects the candidate co-occurrence rules by further combining at least one of one or more first and second words with the multiple words included in the candidate co-occurrence rules according to the related words.

[0023] Such a document analysis support device modifies the co-occurrence rule candidates in accordance with the related words that are further related to the unimportant sentences stored in the unimportant sentence storage unit, and therefore can more appropriately determine unimportance.

[0024] A document analysis support method according to another aspect of the present invention is a document analysis support method executed by a document analysis support device including a display unit, a co-occurrence rule storage unit that stores one or more combined words as co-occurrence rules, an important sentence storage unit that stores one or more first sentences as important sentences, and an unimportant sentence storage unit that stores one or more second sentences different from the first sentences as unimportant sentences, the document analysis support method ... and an unimportant sentence storage unit that stores one or more unimportant sentences. The system includes an importance determination step, a display processing step for displaying important sentences determined to be unimportant in the importance determination step on the display unit, an input step for accepting input of multiple words as candidates for co-occurrence rules for the important sentences displayed on the display unit, and a non-importance determination step for determining whether or not a non-important sentence stored in the non-important sentence storage unit is important based on the candidate co-occurrence rules accepted in the input step, and the display processing step displays the candidate co-occurrence rules accepted in the input step and the determination result that the non-importance sentence is important on the display unit.

[0025] In this document analysis support method, when an important sentence is determined to be unimportant based on the co-occurrence rules stored in the co-occurrence rule storage unit, the important sentence determined to be unimportant is displayed on the display unit, so that a candidate co-occurrence rule for the important sentence determined to be unimportant can be input while referring to the important sentence displayed on the display unit. The document analysis support method determines whether or not a non-important sentence stored in the non-important sentence storage unit is important based on the candidate co-occurrence rules, and displays the candidate co-occurrence rules and the determination result of the non-importance on the display unit. Therefore, the candidate co-occurrence rules can be adjusted while referring to the determination result so that a certain sentence is not determined to be important. Therefore, the document analysis support method can appropriately modify the co-occurrence rules.

[0026] Another aspect of the present invention provides a document analysis support program that causes a computer to function as a co-occurrence rule memory unit that stores one or more combined words as co-occurrence rules, an important sentence memory unit that stores one or more first sentences as important sentences, a non-important sentence memory unit that stores one or more second sentences different from the first sentences as non-important sentences, an importance judgment unit that determines whether an important sentence stored in the important sentence memory unit is important based on the co-occurrence rules stored in the co-occurrence rule memory unit, a display processing unit that displays important sentences that are determined to be unimportant by the importance judgment unit on a display unit, an input unit that accepts input of multiple words as candidates for co-occurrence rules for the important sentences displayed on the display unit, and a non-importance judgment unit that determines whether a non-important sentence stored in the non-important sentence memory unit is important based on the candidate co-occurrence rules accepted by the input unit, and the display processing unit displays the candidate co-occurrence rules accepted by the input unit and the judgment result of the non-importance judgment unit that a sentence is important on the display unit.

[0027] Such a document analysis support program can appropriately modify the co-occurrence rules.

[0028] a display processing unit that displays the target sentence on the display unit when the importance judgment unit judges that the target sentence is not important; an input unit that accepts input of a plurality of words as candidates for co-occurrence rules for the target sentence displayed on the display unit; and an unimportance judgment unit that judges whether the unimportant sentence stored in the unimportant sentence memory unit is important or not, based on the candidate co-occurrence rules accepted by the input unit. The display processing unit displays the candidate co-occurrence rules accepted by the input unit and the judgment result of the unimportance judgment unit that the target sentence is not important on the display unit.

[0029] When a target sentence is determined to be unimportant based on the co-occurrence rules stored in the co-occurrence rule storage unit, this document analysis support device displays the target sentence determined to be unimportant on the display unit, so that the user can input a candidate co-occurrence rule to determine the target sentence as important while referring to the target sentence displayed on the display unit.The document analysis support device determines whether a non-important sentence stored in the non-important sentence storage unit is important based on the candidate co-occurrence rules, and displays the candidate co-occurrence rules and the determination result on the display unit.The document analysis support device can adjust the candidate co-occurrence rules so that a certain sentence is not determined to be unimportant while referring to the determination result.Therefore, the document analysis support device can appropriately modify the co-occurrence rules.

[0030] In another aspect, in the document analysis support device described above, the input unit further includes a synonym-registration unit that receives input of synonymous words that are regarded as synonyms for words in a co-occurrence rule and stores the synonymous words received by the input unit in the co-occurrence rule storage unit in association with the words, and a common word extraction unit that extracts words and synonymous words that are common among a plurality of co-occurrence rules when the co-occurrence rule storage unit stores a plurality of co-occurrence rules, and the display processing unit further displays the common words and synonymous words extracted by the common word extraction unit on the display unit. Preferably, in the document analysis support device described above, the common word extraction unit further extracts at least one of the words and synonymous words associated with the common words and synonymous words extracted by the common word extraction unit.

[0031] In such a document analysis support device, when the co-occurrence rule storage unit stores multiple co-occurrence rules, it extracts words that are common among the multiple co-occurrence rules and words that are considered to be synonyms and displays them on the display unit, so that when correcting the ontology, the correction can be made efficiently by referring to this, and when input of words that are considered to be synonyms is accepted from multiple users, synonymous words entered by other users are also extracted and displayed, so that synonymous words entered by other users can also be referred to.

[0032] In another aspect, in the document analysis support device, the common word extraction unit further extracts, as the common words and synonymous words, words and synonymous words whose word edit distance between a plurality of co-occurrence rules is equal to or less than a threshold. Preferably, in the document analysis support device, the word edit distance is a Levenshtein distance calculated as the minimum number of steps required to transform two character strings into another character string by inserting, deleting, or substituting one character between the two character strings.

[0033] Such a document analysis support device extracts words and synonymous words whose word edit distance is equal to or less than a threshold value as the common words and synonymous words, and can therefore support a user in noticing synonymous words that the same user would not have noticed otherwise.

[0034] In another aspect, in the above-mentioned document analysis support device, the display processing unit displays the common words and synonymous words extracted by the common word extraction unit on the display unit in a graph in which the words and synonymous words are represented by nodes and the associations between the words and synonymous words are represented by edges.

[0035] Such a document analysis support device displays the common words and synonymous words extracted by the common word extraction unit in a graph (e.g., an undirected graph) on a display unit, allowing the relationship between words and synonymous words to be visually grasped and recognized efficiently.

[0036] In another aspect, in the document analysis support device, the display processor further attaches the word edit distance to the edge, or changes a display mode of the edge in accordance with the word edit distance. Preferably, in the document analysis support device, the display mode of the edge is any one of a thickness of the edge, a display color of the edge, and a shade of the display color of the edge.

[0037] Such a document analysis support device attaches the word edit distance to an edge or changes the display mode of the edge depending on the word edit distance, so that the degree (strength) of the association between a word and a word considered to be synonymous can be visually grasped and recognized efficiently. [Effects of the Invention]

[0038] The document analysis support device, the document analysis support method, and the document analysis support program according to the present invention can appropriately correct the rules. [Brief explanation of the drawings]

[0039] [Figure 1]1 is a block diagram showing a configuration of a document analysis support device according to an embodiment. [Figure 2] FIG. 10 is a diagram illustrating an example of a co-occurrence rule. [Figure 3] FIG. 1 is a diagram for explaining an ontology as an example. [Figure 4] FIG. 10 is a diagram for explaining an important document as an example. [Figure 5] FIG. 10 is a diagram illustrating an example of a non-important document. [Figure 6] FIG. 10 is a diagram showing an example of a determination result and co-occurrence rule input screen. [Figure 7] FIG. 10 is a diagram showing an example of a determination result and a co-occurrence rule input screen that displays non-important sentences. [Figure 8] 4 is a flowchart showing the operation of the document analysis support device. [Figure 9] FIG. 10 is a diagram showing an example of a synonymous word processing screen. [Figure 10] 10A and 10B are diagrams illustrating examples of input of synonymous words and graphs, as an example. [Figure 11] FIG. 10 is a diagram for explaining graphs in the third and fourth modified embodiments as an example. DETAILED DESCRIPTION OF THE INVENTION

[0040] Hereinafter, one or more embodiments of the present invention will be described with reference to the drawings. However, the scope of the invention is not limited to the disclosed embodiments. In addition, components with the same reference numerals in each drawing indicate the same components, and their description will be omitted as appropriate. In this specification, when referring to a general term, a reference numeral without a subscript is used, and when referring to an individual component, a reference numeral with a subscript is used.

[0041] A document analysis support device according to an embodiment is a device for supporting document analysis, and in this embodiment, particularly supports the modification (updating) of co-occurrence rules used in the analysis. This document analysis support device includes a display unit, a co-occurrence rule storage unit that stores one or more combined words as co-occurrence rules, an important sentence storage unit that stores one or more first sentences as important sentences, a non-important sentence storage unit that stores one or more second sentences different from the first sentences as non-important sentences, an importance determination unit that determines whether an important sentence stored in the important sentence storage unit is important based on the co-occurrence rules stored in the co-occurrence rule storage unit, a display processing unit that displays an important sentence determined to be unimportant by the importance determination unit on the display unit, an input unit that accepts input of multiple words as candidates for co-occurrence rules for the important sentence displayed on the display unit, and a non-importance determination unit that determines whether an unimportant sentence stored in the non-important sentence storage unit is important based on the candidate co-occurrence rules accepted by the input unit. The display processing unit then displays the co-occurrence rule candidates received by the input unit and the results of the unimportance determination unit determining that the co-occurrence rules are important on the display unit. Such a document analysis support device will be described in more detail below.

[0042] FIG. 1 is a block diagram showing the configuration of a document analysis support device in an embodiment. FIG. 2 is a diagram for explaining co-occurrence rules as an example. FIG. 3 is a diagram for explaining an ontology as an example. FIG. 4 is a diagram for explaining an important document as an example. FIG. 5 is a diagram for explaining an unimportant document as an example. FIG. 6 is a diagram showing a determination result and co-occurrence rule input screen as an example. FIG. 7 is a diagram showing a determination result and co-occurrence rule input screen displaying unimportant sentences as an example.

[0043] A document analysis support device D in the embodiment includes a control processing unit 1, an input unit 2, a display unit 3, an interface unit (IF unit) 4, and a storage unit 5, as shown in FIG.

[0044] The input unit 2 is connected to the control processing unit 1 and is a device that inputs various commands, such as a command to start document analysis and a command to modify (update) co-occurrence rules, as well as various data required to operate the document analysis support device D, such as document names and co-occurrence rule candidates, to the document analysis support device D, and is, for example, a plurality of input switches to which predetermined functions are assigned, a keyboard, a mouse, etc. The display unit 3 is connected to the control processing unit 1 and is a device that outputs the commands and data input from the input unit 2 and the judgment results, etc., under the control of the control processing unit 1, and is, for example, a display device such as a CRT display, a liquid crystal display, or an organic EL display.

[0045] The input unit 2 and the display unit 3 may form a so-called touch panel. In this touch panel, the input unit 2 is a position input device, such as a resistive film type or a capacitive type, that detects an operation position and inputs the position. In this touch panel, the position input device is provided on the display surface of the display device, and one or more input content candidates that can be input are displayed on the display device. When a user touches a display position that displays the input content they want to input, the position is detected by the position input device, and the display content displayed at the detected position is input to the document analysis support device D as the user's operation input content. With such a touch panel, the user can easily intuitively understand the input operation, and therefore a document analysis support device D that is easy for the user to use is provided.

[0046] The IF unit 4 is connected to the control processing unit 1 and is a circuit that inputs and outputs data to and from external devices under the control of the control processing unit 1, and is, for example, an interface circuit for RS-232C, which is a serial communication method, an interface circuit using the Bluetooth (registered trademark) standard, an interface circuit for infrared communication such as the IrDA (Infrared Data Association) standard, and an interface circuit using the USB (Universal Serial Bus) standard. The IF unit 4 is also a circuit that communicates with external devices, and may be, for example, a data communication card or a communication interface circuit conforming to the IEEE802.11 standard or the like.

[0047] The storage unit 5 is connected to the control processing unit 1 and is a circuit that stores various predetermined programs and various predetermined data under the control of the control processing unit 1. The various predetermined programs include, for example, a control processing program, which controls the components 2 to 5 of the document analysis support device D, an importance determination program that determines whether an important sentence stored in the important sentence storage unit 54 (described later) is important based on co-occurrence rules stored in the co-occurrence rule storage unit 52 (described later), a display processing program that displays important sentences determined to be unimportant by the importance determination program on the display unit 3, and an unimportance determination program that determines whether an unimportant sentence stored in the unimportant sentence storage unit 55 (described later) is important based on candidate co-occurrence rules received by the input unit 2. The various predetermined data include, for example, data required to execute each of these programs, such as the document to be analyzed, the co-occurrence rules, the ontology, the important sentences, and the unimportant sentences. The storage unit 5 includes, for example, a ROM (Read Only Memory), which is a nonvolatile storage element, and an EEPROM (Electrically Erasable Programmable Read Only Memory), which is a rewritable nonvolatile storage element. The storage unit 5 also includes a RAM (Random Access Memory), which serves as a working memory for the control processing unit 1 and stores data generated during execution of the predetermined program. The storage unit 5 may also be configured with a hard disk drive having a relatively large storage capacity.

[0048] The storage unit 5 functionally comprises a document storage unit 51, a co-occurrence rule storage unit 52, an ontology storage unit 53, an important sentence storage unit 54, an unimportant sentence storage unit 55, and an erroneously determined sentence storage unit 56.

[0049] The document storage unit 51 stores documents for which analysis support is required. The documents for which analysis support is required may be any type of document, such as a contract, a specification, or an instruction manual. The documents for which analysis support is required are appropriately selected by a user (operator). The documents for which analysis support is required are, for example, input through the input unit 2 and stored in the document storage unit 51 in association with a document ID (document name), which is an identifier for specifying and identifying the document, such as a document name. Alternatively, the documents for which analysis support is required are read, for example, from a storage medium (e.g., a USB memory or an SD card (registered trademark)) on which the documents for analysis support are stored via the IF unit 4, and stored in the document storage unit 51 in association with the document ID. Alternatively, the documents for which analysis support is required are read, for example, from a storage medium (e.g., a CD-R or a DVD-R) on which the documents for analysis support are recorded via a drive device (not shown) connected to the control processing unit 1 and driven under the control of the control processing unit 1, and stored in the document storage unit 51 in association with the document ID. Alternatively, for example, the document is read from a server device that manages the document to be analyzed via the IF unit 4, and stored in the document storage unit 51 in association with the document ID.

[0050] The co-occurrence rule storage unit 52 stores one or more co-occurrence rules, each consisting of a combination of multiple words. The co-occurrence occurs when a certain word appears in a sentence and another limited word frequently appears in the sentence. Therefore, by combining the certain word with the other limited word to form a rule, the sentence can be extracted. The co-occurrence rules are appropriately set by the user. The co-occurrence rules are stored in the co-occurrence rule storage unit 52 in a table format, as shown in FIG. 2, for example. The co-occurrence rule registration table RT shown in FIG. 2 for registering the co-occurrence rules includes a rule number field 61 for registering a rule number (rule number), which is an identifier for specifying and identifying the co-occurrence rule, and a co-occurrence rule field 62 for registering the co-occurrence rule corresponding to the rule number registered in the rule number field 61. The table has a number of records corresponding to the number of co-occurrence rules. 2, for example, in a record in which "1" is registered in the rule No. field 61, "gear, material, stainless, steel" is registered as a co-occurrence rule in the co-occurrence rule field 62. The co-occurrence rules are input via the input unit 2, the IF unit 4, or a drive device (not shown), in the same way as the document to be analyzed, and are stored in the co-occurrence rule storage unit 52.

[0051] The ontology storage unit 53 stores an ontology that represents relationships between words. This ontology is set appropriately by the user. The ontology is stored in the ontology storage unit 53 in a tree structure, for example, as shown in FIG. 3. In the ontology tree structure OG representing the ontology shown in FIG. 3, related words are sequentially connected by parent-child nodes in each layer LAn starting from the root node of the first layer LA1. For example, the root node (parent node) "material" of the first layer LA1 is connected to "steel" and "copper" as child nodes in the next layer, the second layer LA2, and "material" is related (associated) with "steel" and "copper", respectively. The parent node "steel" of the second layer LA2 is connected to "stainless steel" as a child node in the next layer, the third layer LA3, and "steel" is related (associated) with "stainless steel". The parent node "stainless steel" at the third level LA3 is connected to "ss" as a child node at the next level, the fourth level LA4, and "stainless steel" is related (associated) with "ss." In an ontology, the strength of the relationship between words (degree of association) is represented by the hierarchical distance between the words. In the above example, "steel" is closest to "material" at the first level and has the strongest (highest) degree of association, while "ss" is farthest from "material" at the third level and has the weakest (lowest) degree of association. The ontology, like the document to be analyzed, is input via the input unit 2, the IF unit 4, or a drive device (not shown) and stored in the ontology storage unit 53.

[0052] The important sentence storage unit 54 stores one or more first sentences as important sentences. These important sentences are set appropriately by the user. The important sentences are stored in the important sentence storage unit 54 in table format, for example, as shown in FIG. 4. The important sentence registration table DL shown in FIG. 4 for registering important sentences includes an important sentence field 63 for registering important sentences and a Comments field 64 for registering comments corresponding to the important sentences registered in the important sentence field 63, and includes a number of records corresponding to the number of important sentences. The comments may be free-form and arbitrary. The important sentences, like the documents to be analyzed, are input via the input unit 2, the IF unit 4, or a drive device (not shown) and stored in the important sentence storage unit 54.

[0053] The non-important sentence storage unit 55 stores one or more second sentences different from the first sentence (important sentence) as non-important sentences. The non-important sentences are set appropriately by the user. The non-important sentences are stored in the non-important sentence storage unit 55 in a table format, for example, as shown in FIG. 5. The non-important sentence registration table NL shown in FIG. 5 for registering non-important sentences includes a non-important sentence field 65 for registering non-important sentences, and includes records in the number corresponding to the number of non-important sentences. The non-important sentences are input via the input unit 2, the IF unit 4, or a drive device (not shown) and stored in the non-important sentence storage unit 55, just like the document to be analyzed.

[0054] As described above, the document may be any type of document. For example, if the document is a contract, the important text is a text included in the contract and designated as important, the unimportant text is a text included in the contract and designated as unimportant, and the comment is a response to or rejection of a clause (provision) included in the contract or an alternative clause (alternative clause). Alternatively, if the document is a specification, the important text is a text included in the specification and designated as important, the unimportant text is a text included in the specification and designated as unimportant, and the comment is a response to or rejection of a specification included in the specification, or an alternative, if the specification is a requirements specification for a customer's product. Alternatively, if the document is an instruction manual, the important text is a text included in the instruction manual and designated as important, the unimportant text is a text included in the instruction manual and designated as unimportant, and the comment is an annotation or reference to an explanation included in the instruction manual.

[0055] The incorrectly determined sentence storage unit 56 stores important sentences that have been determined to be unimportant by the importance determination program.

[0056] The control processing unit 1 is a circuit that controls each of the units 2 to 5 of the document analysis support device D according to the function of each unit, supports document analysis, and supports the modification (updating) of co-occurrence rules used in the analysis. The control processing unit 1 is configured, for example, with a CPU (Central Processing Unit) and its peripheral circuits. By executing a control processing program, the control processing unit 1 functionally comprises a control unit 11, an importance determination unit 12, a display processing unit 13, and an unimportance determination unit 14.

[0057] The control unit 11 controls each of the units 2 to 5 of the document analysis support device D according to the function of each unit, and is responsible for overall control of the document analysis support device D. The control unit 11 stores input or read documents, co-occurrence rules, ontologies, important sentences, and unimportant sentences in a document storage unit 51, a co-occurrence rule storage unit 52, an ontology storage unit 53, an important sentence storage unit 54, and an unimportant sentence storage unit 55, respectively.

[0058] The importance determination unit 12 first determines whether each sentence in a document stored in the document storage unit 51 is important based on the co-occurrence rules stored in the co-occurrence rule storage unit 52. More specifically, the importance determination unit 12 determines that the sentence to be analyzed is important if the sentence included in the document (the sentence to be analyzed) contains all of the multiple words combined in the co-occurrence rule, and determines that the sentence to be analyzed is not important if the sentence to be analyzed does not contain some or all of the multiple words combined in the co-occurrence rule. If the co-occurrence rule storage unit 52 stores multiple combinations of multiple words as co-occurrence rules, the importance determination unit 12 determines whether each of the multiple combinations is important for the sentence to be analyzed. For example, if the co-occurrence rule memory unit 52 stores a first co-occurrence rule of rule No. 1 and a second co-occurrence rule of rule No. 2, the importance determination unit 12 determines whether the sentence to be analyzed is important based on the first co-occurrence rule, and determines whether the sentence to be analyzed is important based on the second co-occurrence rule.

[0059] In this embodiment, since ontologies are stored in the ontology storage unit 53, the importance determination unit 12 determines whether each sentence in a document stored in the document storage unit 51 is important based on the co-occurrence rules stored in the co-occurrence rule storage unit 52 and the ontologies stored in the ontology storage unit 53. More specifically, if there is a word related in the ontology among multiple words combined in a co-occurrence rule, the importance determination unit 12 replaces the word in the co-occurrence rule with the word related in the ontology, and determines whether the sentence to be analyzed is important based on the co-occurrence rule after the replacement. If there are multiple co-occurrence rules after the replacement, the importance determination unit 12 determines that the sentence to be analyzed is important if any of the co-occurrence rules after the replacement determines that the sentence is important. For example, since "material" in the co-occurrence rule "gear, material, stainless, steel" of rule No. "1" is related to "steel," "copper," "stainless steel," and "ss" respectively in the ontology, the importance determination unit 12 determines whether the sentence to be analyzed is important based on "gear, material, stainless, steel," determines whether the sentence to be analyzed is important based on "gear, steel, stainless, steel," determines whether the sentence to be analyzed is important based on "gear, copper, stainless, steel," determines whether the sentence to be analyzed is important based on "gear, stainless steel, stainless, steel," determines whether the sentence to be analyzed is important based on "gear, ss, stainless, steel," and determines whether the sentence to be analyzed is important if it is determined to be important by the co-occurrence rule before replacement, or if it is determined to be important by any of the co-occurrence rules after replacement.

[0060] In the above description, all levels LAn were used, but the levels LAn used in ontology replacement may be limited, such as from the first level LA1 to the second level LA2, or from the first level LA1 to the third level LA3, etc.

[0061] The display processing unit 13 firstly displays the determination result of the importance determination unit 12 on the display unit 3. More specifically, the display processing unit 13 displays each sentence of the document stored in the document storage unit 51 on the display unit 3 so that the display mode of a sentence determined to be important by the importance determination unit 12 is different from the display mode of a sentence determined to be unimportant by the importance determination unit 12. More specifically, the display processing unit 13 highlights words corresponding to co-occurrence rules in the sentences determined to be important by the importance determination unit 12, and displays each sentence of the document stored in the document storage unit 51 on the display unit 3. Instead of highlighting, the display color of the words may be changed.

[0062] In this embodiment, the importance determination unit 12 secondly determines whether an important sentence stored in the important sentence storage unit 54 is important based on the co-occurrence rules stored in the co-occurrence rule storage unit 52. More specifically, as in the first case described above, the importance determination unit 12 determines that an important sentence to be analyzed is important if the important sentence stored in the important sentence storage unit 54 (the important sentence to be analyzed) contains all of the multiple words combined in the co-occurrence rule, and determines that the important sentence to be analyzed is not important if the important sentence to be analyzed does not contain some or all of the multiple words combined in the co-occurrence rule. The important sentence to be analyzed that is determined to be not important is stored in the incorrectly determined sentence storage unit 56. If the co-occurrence rule storage unit 52 stores multiple combinations of multiple words as co-occurrence rules, the importance determination unit 12 determines whether each of the multiple combinations is important for the important sentence to be analyzed. When there are multiple important sentences stored in the important sentence storage unit 54, the importance determination unit 12 determines whether each of the multiple important sentences is important or not based on the co-occurrence rules stored in the co-occurrence rule storage unit 52. In this embodiment, the importance determination unit 12 determines whether each important sentence stored in the important sentence storage unit 54 is important or not based on the co-occurrence rules stored in the co-occurrence rule storage unit 52 and the ontologies stored in the ontology storage unit 53.

[0063] Secondly, in this embodiment, the display processing unit 13 displays on the display unit 3 important sentences that are determined to be unimportant by the importance determination unit 12 from among the important sentences stored in the important sentence storage unit 54.

[0064] The input unit 2 accepts input of multiple words as candidates for co-occurrence rules for important sentences stored in the important sentence storage unit 54 that have been determined to be unimportant by the importance determination unit 12 and displayed on the display unit 3 by the display processing unit 13.

[0065] The unimportance determination unit 14 determines whether a non-important sentence stored in the unimportant sentence storage unit 55 is important based on the candidate co-occurrence rules received by the input unit 2. More specifically, the unimportance determination unit 14 determines that a non-important sentence to be analyzed is important if the non-important sentence (the non-important sentence to be analyzed) stored in the unimportant sentence storage unit 55 contains all of the multiple words combined as the candidate co-occurrence rules, and determines that the non-important sentence to be analyzed is unimportant if the non-important sentence to be analyzed does not contain some or all of the multiple words combined as the candidate co-occurrence rules. The non-important sentence to be analyzed that is determined to be unimportant is stored in the storage unit 5. When there are multiple unimportant sentences stored in the unimportant sentence storage unit 55, the unimportance determination unit 14 determines whether each of the multiple unimportant sentences is important based on the candidate co-occurrence rules.

[0066] Thirdly, in this embodiment, the display processing unit 13 displays the candidates for co-occurrence rules received by the input unit 2 and the determination result of the unimportance determination unit 14 determining that the sentences are unimportant on the display unit 3. The determination result includes the number of unimportant sentences determined to be important by the unimportance determination unit 14. The number of unimportant sentences to be analyzed that are stored in the memory unit 5 and that are determined to be unimportant is counted, and the count result is displayed on the display unit 3.

[0067] In this embodiment, fourthly, when displaying the judgment result, the display processing unit 13 displays an operation button on the display unit 3 for displaying the non-important sentences judged to be important by the non-importance judgment unit 14 on the display unit 3, and when the operation button is operated, displays the non-important sentences judged to be important by the non-importance judgment unit 14 on the display unit 3. The non-important sentences to be analyzed that are judged to be unimportant and stored in the memory unit 5 are displayed on the display unit 3.

[0068] Fifthly, in this embodiment, the display processing unit 13 displays on the display unit 3 the relationships between multiple words as candidates for co-occurrence rules received by the input unit 2 based on the ontology stored in the ontology storage unit 53.

[0069] For example, the display processing unit 13 displays on the display unit 3 the judgment results shown in Figure 6, the important sentences judged to be unimportant by the importance judgment unit 12, the candidate co-occurrence rules received by the input unit 2, and the judgment results judged to be important by the unimportance judgment unit 14, etc.

[0070] 6 includes a first display area 71 displaying important sentences determined to be unimportant by the importance determination unit 12, a second display area 72 displaying comments corresponding to important sentences determined to be unimportant by the importance determination unit 12 and displayed in the first display area 71 in the important sentence registration table DL, a third display area 73 displaying co-occurrence rule candidates received by the input unit 2, a fourth display area 74 displaying inter-word relationships for multiple words as co-occurrence rule candidates received by the input unit 2 and displayed in the third display area 73 based on the ontology stored in the ontology storage unit 53, a fifth display area 75 displaying the number of unimportant sentences determined to be important by the unimportance determination unit 14, and a confirm button 76 for inputting an instruction to confirm the co-occurrence rule candidates displayed in the third display area 73 as co-occurrence rules. If the input unit 2 does not receive input of co-occurrence rule candidates, the third and fourth display areas 73 and 74 are left blank. If, based on the ontology, there are no words related to the multiple words as candidate co-occurrence rules displayed in the third display area 73, the fourth display area 74 is left blank. When the Confirm button 76 is operated using the input unit 2, the control unit 11 confirms the candidate co-occurrence rules displayed in the third display area 73 as co-occurrence rules, assigns a rule number to the confirmed co-occurrence rule, registers the confirmed co-occurrence rule in the co-occurrence rule registration table RT in association with the assigned rule number, and stores the registered co-occurrence rule in the co-occurrence rule storage unit 52.

[0071] In this embodiment, the fifth display area 75 also serves as an operation button for displaying unimportant sentences determined to be important by the unimportance determination unit 14 on the display unit 3. When the fifth display area 75 is operated by using the input unit 2 for input, for example, as shown in Fig. 7 , a window 77 for displaying unimportant sentences determined to be important by the unimportance determination unit 14 is opened, and the unimportant sentences determined to be unimportant by the unimportance determination unit 14 are displayed in this window 77.

[0072] Then, the input unit 2 accepts changes to the multiple words as candidate co-occurrence rules that have already been accepted by the input unit 2. Secondly, in this embodiment, the unimportance determination unit 14 re-determines whether or not the unimportant sentences stored in the unimportant sentence storage unit 55 are important based on the changed co-occurrence rule candidates accepted by the input unit 2. Sixthly, in this embodiment, the display processing unit 13 displays on the display unit 3 the changed co-occurrence rule candidates accepted by the input unit 2 and the re-determination result that the unimportance determination unit 14 has determined that the unimportance rule candidates are important. For example, the determination result, co-occurrence rule input screen SC shown in FIG. 6 is updated with the changed co-occurrence rule candidates, the re-determination result, etc.

[0073] The importance determination unit 12 determines whether or not an important sentence in the important sentence storage unit 54 is important, and the unimportance determination unit 14 determines whether or not an unimportant sentence in the unimportant sentence storage unit 55 is important. The importance determination unit 12 and the unimportance determination unit 14 only differ in the sentences they determine as important, but use the same logic for determining whether or not a sentence is important, and both determine that a sentence is important if a co-occurrence rule matches the sentence they determine as important.

[0074] The control processing unit 1, input unit 2, display unit 3, IF unit 4 and storage unit 5 can be configured by, for example, a desktop, notebook or tablet computer.

[0075] Next, the operation of this embodiment will be described with reference to the flowchart of FIG.

[0076] When the document analysis support device D having such a configuration is turned on, it initializes the necessary parts and starts its operation. By executing the control processing program, the control processing unit 1 is functionally configured with a control unit 11, an importance determination unit 12, a display processing unit 13, and an unimportance determination unit 14.

[0077] First, in document analysis support, a document to be analyzed, a co-occurrence rule, and an ontology are input and stored in the document storage unit 51, the co-occurrence rule storage unit 52, and the ontology storage unit 53, respectively, in the storage unit 5. When an instruction to start document analysis is given, the document analysis support device D causes the importance determination unit 12 of the control processing unit 1 to determine whether each sentence in the document stored in the document storage unit 51 is important based on the co-occurrence rules stored in the co-occurrence rule storage unit 52 and the ontology stored in the ontology storage unit 53. After the determination for each sentence in the document storage unit 51 is completed, the document analysis support device D causes the display processing unit 13 of the control processing unit 1 to display each sentence in the document stored in the document storage unit 51 on the display unit 3 so that the display mode of each sentence in the document stored in the document storage unit 51 is different from the display mode of sentences determined to be important by the importance determination unit 12, and then terminates the process.

[0078] Next, we will explain how to modify co-occurrence rules. A user (operator) prepares a specified document, extracts important and unimportant sentences from the document, and prepares important and unimportant documents. The user then creates co-occurrence rules from the important sentences to prepare template co-occurrence rules, and creates an ontology to prepare the ontology. The template co-occurrence rules and the ontology may be ones that have been created in the past.

[0079] Then, in FIG. 8, first, the document analysis support device D receives input of co-occurrence rules (the template co-occurrence rules) via the control processing unit 1 through the input unit 2, IF unit 4, etc., and stores them in the co-occurrence rule memory unit 52 of the memory unit 5 (S1).

[0080] Next, the document analysis support device D receives an ontology input via the input unit 2, IF unit 4, etc., through the control processing unit 1, and stores it in the ontology storage unit 53 of the storage unit 5 (S2).

[0081] Next, the document analysis support device D receives input of important sentences through the input unit 2, IF unit 4, etc. by the control processing unit 1, and stores them in the important sentence storage unit 54 of the storage unit 5 (S3).

[0082] Next, the document analysis support device D receives input of unimportant sentences through the control processing unit 1 via the input unit 2, IF unit 4, etc., and stores them in the unimportant sentence storage unit 55 of the storage unit 5 (S4).

[0083] Next, the document analysis support device D determines whether or not the important sentences stored in the important sentence storage unit 54 (important sentences to be analyzed) are important based on the co-occurrence rules stored in the co-occurrence rule storage unit 52, using the importance determination unit 12 of the control processing unit 1, and stores the important sentences to be analyzed that are determined to be unimportant in the incorrectly determined sentence storage unit 56 (S5). In this embodiment, when making the determination, the importance determination unit 12 determines whether or not the important sentences stored in the important sentence storage unit 54 are important based on the co-occurrence rules stored in the co-occurrence rule storage unit 52 and the ontologies stored in the ontology storage unit 53.

[0084] Next, the document analysis support device D determines whether or not the importance determination section 12 of the control processing section 1 has completed the importance determination for all of the important sentences stored in the important sentence storage section 54 (S6). If the result of this determination is that the determination has been completed for all of the important sentences stored in the important sentence storage section 54 (Yes), the document analysis support device D then performs process S7. On the other hand, if the result of the above determination is that the determination has not been completed for all of the important sentences stored in the important sentence storage section 54 (No), the document analysis support device D returns the process to process S5. Therefore, process S5 is repeatedly performed until the importance determination has been completed for all of the important sentences stored in the important sentence storage section 54.

[0085] In process S7, the document analysis support device D displays, for example, the determination result shown in FIG. 6 on the co-occurrence rule input screen SC in the first and second display areas 71 and 72, respectively, important sentences determined to be unimportant by the importance determination unit 12 and corresponding comments for the important sentences stored in the important sentence storage unit 54, and accepts input of co-occurrence rule candidates. When an input operation is performed on the fifth display area 75 using the input unit 2, a window 77 is opened, and unimportant sentences determined to be important by the unimportance determination unit 14 are displayed in this window 77. The user refers to the important sentences determined to be unimportant by the importance determination unit 12 and the corresponding comments displayed in the first and second display areas 71 and 72, respectively, and inputs co-occurrence rule candidates from the input unit 2. When a user inputs a candidate co-occurrence rule using the input unit 2, the display processing unit 13 displays the candidate co-occurrence rule in the third display area 73, and displays the relationships between the multiple words as candidate co-occurrence rules displayed in the third display area 73 in the fourth display area 74 based on the ontology stored in the ontology memory unit 53.

[0086] When the co-occurrence rule candidates are input, the document analysis support device D then causes the unimportance determination unit 14 of the control processing unit 1 to determine whether the unimportant sentences stored in the unimportant sentence storage unit 55 are important or not, based on the co-occurrence rule candidates received by the input unit 2, and causes the display processing unit 13 to display on the display unit 3 the determination result that the unimportance determination unit 14 has determined that the sentences are important (S8). For example, the number of unimportant sentences determined to be important by the unimportance determination unit 14 is displayed as the determination result in the fifth display area 75 of the co-occurrence rule input screen SC.

[0087] Next, the document analysis support device D determines whether the candidate co-occurrence rule is confirmed by the control unit 11 of the control processing unit 1 (S9). If the result of this determination is that the candidate co-occurrence rule is confirmed (Yes), the document analysis support device D then performs process S10. For example, the user refers to the determination result, that is, the number of unimportant sentences determined to be important by the unimportance determination unit 14, displayed in the fifth display area 75 of the co-occurrence rule input screen SC, and determines whether to confirm the candidate co-occurrence rule displayed in the third display area 73 as a co-occurrence rule. If it is determined that the candidate co-occurrence rule is confirmed, the user operates the Confirm button 76 using the input unit 2. Upon receiving the input operation of the Confirm button 76, the control unit 11 determines that the candidate co-occurrence rule is confirmed (Yes). On the other hand, if the result of this determination is that the candidate co-occurrence rule is not confirmed (No), the document analysis support device D returns the process to process S7. For example, if the control unit 11 does not accept an input operation of the Confirm button 76 within a predetermined time after displaying on the display unit 3 the determination results (here, the number) that the unimportance determination unit 14 has determined to be important, the control unit 11 determines that the results are not confirmed (No). Alternatively, for example, an Unconfirm button for inputting an instruction not to confirm the co-occurrence rule candidates displayed in the third display area 73 as co-occurrence rules may be further provided on the determination result / co-occurrence rule input screen SC. In this case, the control unit 11 determines that the results are not confirmed (No) after displaying on the display unit 3 the determination results (here, the number) that the unimportance determination unit 14 has determined to be important (here, the number) and then accepts an input operation of the Unconfirm button.

[0088] By returning the process to process S7, in process S7, the input unit 2 accepts changes to the multiple words as candidate co-occurrence rules that have already been accepted by the input unit 2, and in the subsequent process S8, the unimportance determination unit 14 re-determines whether the unimportant sentences stored in the unimportant sentence storage unit 55 are important or not based on the candidate co-occurrence rules after the change accepted by the input unit 2, and the display processing unit 13 displays the candidate co-occurrence rules after the change accepted by the input unit 2 and the re-determination result that the unimportance determination unit 14 has judged them to be important on the display unit 3. Processes S7 to S9 are repeated until the candidate co-occurrence rules are determined.

[0089] Meanwhile, in the process S10, the document analysis support device D determines the candidate co-occurrence rule by the control unit 11 of the control processing unit 1, and modifies (updates) the co-occurrence rule stored in the co-occurrence rule storage unit 52 with the determined candidate co-occurrence rule. In the example described above, the control unit 11 determines the candidate co-occurrence rule displayed in the third display area 73 as a co-occurrence rule, assigns a rule number to the determined co-occurrence rule, and registers the determined co-occurrence rule in the co-occurrence rule registration table RT in association with the assigned rule number, and stores the result in the co-occurrence rule storage unit 52.

[0090] Next, the document analysis support device D determines, by the control unit 11 of the control processing unit 1, whether or not processing has been completed for all of the important sentences determined by the importance determination unit 12 in step S5 as unimportant (S11). That is, the control unit 11 determines whether or not processing has been completed for all of the important sentences stored in the erroneously determined sentence storage unit 56. If the result of this determination is that processing has been completed for all of the important sentences (Yes), the document analysis support device D terminates this processing. On the other hand, if the result of the determination is that processing has not been completed for all of the important sentences (No), the document analysis support device D returns the processing to step S7. Therefore, steps S7 to S11 are repeated until processing has been completed for all of the important sentences stored in the erroneously determined sentence storage unit 56.

[0091] As described above, the document analysis support device D and the document analysis support method and program implemented therein in the embodiment display the unimportant important sentence on the display unit 3 when the co-occurrence rules stored in the co-occurrence rule storage unit 52 determine that the important sentence is unimportant. Therefore, the user can input a candidate co-occurrence rule for the unimportant important sentence while referring to the important sentence displayed on the display unit 3. The document analysis support device D, document analysis support method, and document analysis support program determine whether the unimportant sentence stored in the unimportant sentence storage unit 55 is important based on the candidate co-occurrence rules, and display the candidate co-occurrence rules and the result of the determination on the display unit 3. Therefore, the user can adjust the candidate co-occurrence rules so that a certain sentence is not determined to be important while referring to the result of the determination. Therefore, the document analysis support device D, document analysis support method, and document analysis support program can appropriately modify the co-occurrence rules. The document analysis support device D, document analysis support method, and document analysis support program can extract co-occurrence rules that were not included in the co-occurrence rules of the template due to, for example, user oversight, and can correct the co-occurrence rules of the template.

[0092] The document analysis support device D, document analysis support device method, and document analysis support program display the number of unimportant sentences determined to be important by the unimportance determination unit 14, and the co-occurrence rule candidates can be adjusted so that this number becomes 0.

[0093] The document analysis support device D, document analysis support device method, and document analysis support program can display non-important sentences determined to be important by the non-importance determination unit 14 on the display unit 3, and therefore can efficiently adjust the co-occurrence rule candidates by referring to the words contained in the non-important documents so as not to determine the non-important sentences as important (so as not to extract the non-important sentences).

[0094] The document analysis support device D, document analysis support device method, and document analysis support program determine whether something is important or not based on ontology, so even if a word contained in an important sentence is a paraphrased word, it can be appropriately determined to be important.

[0095] The document analysis support device D, document analysis support device method, and document analysis support program display words related to (associated with) multiple words as candidates for co-occurrence rules based on ontology, making it possible to adjust the candidates for co-occurrence rules using the related words.

[0096] The document analysis support device D, document analysis support method, and document analysis support program change the words of the co-occurrence rule candidates and re-evaluate using the changed co-occurrence rule candidates, so that the appropriateness of the changed co-occurrence rule candidates can be recognized and the co-occurrence rule candidates can be efficiently adjusted.

[0097] 1, the document analysis support device D may further include, in the control processing unit 1, a co-occurrence rule modification unit 15 that modifies the co-occurrence rules stored in the co-occurrence rule storage unit 52 in accordance with related words related to the important sentences stored in the important sentence storage unit 54 (first modified embodiment). In this case, the importance determination unit 12 determines whether or not an important sentence stored in the important sentence storage unit 54 is important based on the co-occurrence rules modified by the co-occurrence rule modification unit 15. The document analysis support device D, document analysis support device method, and document analysis support program of the first modified embodiment modify the co-occurrence rules in accordance with related words related to the important sentences stored in the important sentence storage unit 54, thereby enabling more appropriate determination of importance.

[0098] The related words are, for example, first words included in the title of a document including an important sentence stored in the important sentence storage unit 54, or second words related to the first words. The co-occurrence rule modification unit 15 modifies the co-occurrence rules, for example, by deleting some of the words included in the co-occurrence rules in accordance with the related words. Alternatively, the co-occurrence rule modification unit 15 modifies the co-occurrence rules by further combining at least one or more first and second words with the words included in the co-occurrence rules in accordance with the related words.

[0099] For example, in a customer requirement specification for a "reducer," the document title usually includes "reducer," and the specification sentences are sometimes created based on the "reducer." In such a case, "reducer" is set as the first word, and the co-occurrence rule modification unit 15 modifies the co-occurrence rule by deleting "reducer" from the multiple words included in the co-occurrence rule. Alternatively, the co-occurrence rule modification unit 15 modifies the co-occurrence rule by further combining "reducer" with the multiple words included in the co-occurrence rule.

[0100] For example, in a customer requirement specification for a "reducer," a reducer naturally includes a "gear," so the specification sentence may be created assuming "gear." In such a case, "reducer" is set as the first word, "gear" is set as the second word, and the co-occurrence rule modification unit 15 modifies the co-occurrence rule by deleting "gear" from the multiple words included in the co-occurrence rule. Alternatively, the co-occurrence rule modification unit 15 modifies the co-occurrence rule by further combining "gear" with the multiple words included in the co-occurrence rule.

[0101] For example, in a customer requirement specification for "reducer," the specification sentence may be created assuming "reducer" and "gear." In such a case, "reducer" is set as the first word, "gear" is set as the second word, and the co-occurrence rule modification unit 15 modifies the co-occurrence rule by deleting "reducer" and "gear" from the multiple words included in the co-occurrence rule. Alternatively, the co-occurrence rule modification unit 15 modifies the co-occurrence rule by further combining "reducer" and "gear" with the multiple words included in the co-occurrence rule.

[0102] In the above example, the title of the document may include a word that refers to the "reducer," such as the product name or abbreviation of the "reducer," and the word that refers to the "reducer" may be the first word, and "reducer" or "gear" may be the second word.

[0103] In this first modified embodiment, the related words may be further related to unimportant sentences stored in the unimportant sentence storage unit 55, the co-occurrence rule modification unit 15 may modify the co-occurrence rule candidates received by the input unit 2 in accordance with the related words, and the unimportance determination unit 14 may determine whether the unimportant sentences stored in the unimportant sentence storage unit 55 are important based on the modified co-occurrence rule candidates modified by the co-occurrence rule modification unit 15 (second modified embodiment). The document analysis support device D, document analysis support method, and document analysis support program of the second modified embodiment modify the co-occurrence rule candidates in accordance with the related words further related to the unimportant sentences stored in the unimportant sentence storage unit 55, thereby more appropriately determining unimportance. The related words are first words or second words related to the first words included in the title of a document composed of important sentences stored in the important sentence storage unit 54 and unimportant sentences stored in the unimportant sentence storage unit 55. For example, the co-occurrence rule modification unit 15 modifies the candidate co-occurrence rule by deleting some of the words included in the candidate co-occurrence rule in accordance with the related words. Alternatively, for example, the co-occurrence rule modification unit 15 modifies the candidate co-occurrence rule by further combining at least one or more first and second words with the words included in the candidate co-occurrence rule in accordance with the related words.

[0104] Furthermore, in the above-described embodiment, important sentences are stored in the important sentence storage unit 54 before their importance is determined. However, a single important sentence may be acquired from the input unit 2 (one example of a target sentence acquisition unit) or the IF unit 4 (another example of a target sentence acquisition unit) as a target sentence to be analyzed. In this case, the importance determination unit 12 determines whether the target sentence acquired by the input unit 2 or the IF unit 4 is important based on the co-occurrence rules stored in the co-occurrence rule storage unit 52, and the display processing unit 13 displays the target sentence on the display unit 3 if the importance determination unit 12 determines that the target sentence is not important. In this manner, important sentences may be input one by one to the document analysis support device D, and co-occurrence rules for the important sentences may be generated each time, and the co-occurrence rules stored in the co-occurrence rule storage unit 52 may be sequentially modified. With such a document analysis support device D, document analysis support method, and document analysis support program, rules can be appropriately modified in the same manner as with the document analysis support device D, document analysis support method, and document analysis support program of the above-described embodiments.

[0105] In the above-described embodiment and the first and second variations, synonymous words that are regarded as synonyms for the words in the co-occurrence rules may be introduced (third variation).

[0106] In this case, the input unit 2 of the document analysis support device D further accepts input of synonymous words that are considered synonymous with words in the co-occurrence rules. As shown by the dashed lines in FIG. 1 , the document analysis support device D further includes, in the control processing unit 1, a synonymous word registration unit 16 and a common word extraction unit 17. The synonymous word registration unit 16 associates the synonymous words accepted by the input unit 2 with the words and stores them in the co-occurrence rule storage unit 52. When the co-occurrence rule storage unit 52 stores multiple co-occurrence rules, the common word extraction unit 17 extracts words that are common among the multiple co-occurrence rules and words that are considered synonymous. In this third modified embodiment, the common word extraction unit 17 further extracts at least one of the words and words that are considered synonymous that are associated with the common words and words that are considered synonymous extracted by the common word extraction unit 17 as described above. Then, the display processing unit 13 in the control processing unit 1 of the document analysis support device D further displays the common words and synonymous words extracted by the common word extraction unit 17 on the display unit 3. For example, the display processing unit 13 displays the common words and synonymous words extracted by the common word extraction unit 17 on the display unit 3 in a graph (e.g., an undirected graph) in which the words and synonymous words are represented by nodes and the associations between the words and synonymous words are represented by edges.

[0107] Fig. 9 is a diagram showing an example of a synonymous word processing screen. Fig. 10 is a diagram for explaining an example of input of synonymous words and a graph. Figs. 10A to 10E show input examples of synonymous words, respectively, and Fig. 10F shows a graph (undirected graph).

[0108] More specifically, for example, when a command instructing input of synonymous words is received by the input unit 2, the document analysis support device D causes the display processing unit 13 of the control processing unit 1 to display a synonymous word processing screen for performing processing related to synonymous words on the display unit 3. In this third modified embodiment, the synonymous word processing screen is a screen for receiving input of synonymous words to be regarded as synonymous for words in a co-occurrence rule, for example, and for displaying the common words and synonymous words extracted by the common word extraction unit 17 in a graph. As shown in FIG. 9, the synonym-regarding-word processing screen SA includes a synonym-regarding-word display area 81 for displaying words of a co-occurrence rule and corresponding synonym-regarding words received by the input unit 2, a first scroll button 82a for inputting an instruction to scroll the synonym-regarding-word display area 81 in one direction, a second scroll button 82b for inputting an instruction to scroll the synonym-regarding-word display area 81 in the other direction opposite to the one direction, a graph-displaying-word display area 83 for displaying words for which graphs are to be displayed and synonym-regarding words received by the input unit 2, a graph display area 84 for displaying graphs, and a scroll button 82c for inputting an instruction to scroll the graph display area 84 in one direction. The screen SA also includes a third scroll button 85a for inputting a scroll command, a fourth scroll button 85b for inputting an instruction to scroll the graph display area 84 in the opposite direction, a confirm button 86 for inputting an instruction to store the words of the co-occurrence rule and the corresponding synonymous words displayed in the synonymous word display area 81 in the co-occurrence rule storage unit 52, a graph display button 87 for inputting an instruction to display a graph corresponding to the words displayed in the graph display word display area 83 and the corresponding synonymous words in the graph display area 84, and an end button 88 for inputting an instruction to end the synonymous word processing screen SA. The synonymous words are input from the input unit 2 in parentheses following the words of the co-occurrence rule (the words of the co-occurrence rule (synonymous words)), as shown in each of Figures 10A to 10E. The graph is configured with nodes representing the words and the synonymous words and edges representing the associations between the words and the synonymous words.

[0109] 9, the synonym-assigned word display area 81 not only displays the words of a co-occurrence rule received by the input unit 2 and the corresponding synonym-assigned words, but also displays multiple words in the co-occurrence rule received by the input unit 2. That is, the document analysis support device D can receive input of multiple words in a co-occurrence rule at the input unit 2 using the synonym-assigned word processing screen SA shown in FIG. 9 and store them in the co-occurrence rule storage unit 52. Therefore, the co-occurrence rule input screen SC shown in FIGS. 6 and 7 can be used for inputting and storing (registering) synonym-assigned words. That is, when inputting a candidate co-occurrence rule on the co-occurrence rule input screen SC shown in FIGS. 6 and 7, a synonym-assigned word associated with the candidate co-occurrence rule word is input in parentheses following the candidate co-occurrence rule word from the input unit 2, and the confirmation button 76 is pressed, whereby the document analysis support device D can input the synonym-assigned word and store it in the co-occurrence rule storage unit 52.

[0110] For example, when the synonym-regarding word processing screen SA is displayed on the display unit 3, the first user (first operator) uses the input unit 2 to input "23 (voting) (2 out of 3); different (separate) (each); jb (junction box)." This means that "23," "different," and "jb" are multiple words in the co-occurrence rule, "voting" and "2 out of 3" are each synonym-regarding words for "23," "separate" and "each" are each synonym-regarding words for "different," and "junction box" is a synonym-regarding word for "jb." When this input is accepted, "23 (voting) (2 out of 3); different (separate) (each); jb (junction box)" shown in FIG. 10A is displayed in the synonym-regarding word display area 81 of the synonym-regarding word processing screen SA. In this situation, when the first user operates the confirm button 86, the document analysis support device D causes the synonymous word registration unit 16 to store "23 (voting) (2 out of 3); different (sepaate) (each); jb (junction box)," which has been input through the input unit 2 and is displayed in the synonymous word display area 81, in the co-occurrence rule storage unit 52. In this third modified form, similar to the input method and display method, the synonymous words associated with the words of the co-occurrence rule are stored (registered) in the co-occurrence rule storage unit 52 in parentheses following the words of the co-occurrence rule. Then, the first user operates the end button 88 to close the synonymous word processing screen SA.

[0111] 9, by operating the first scroll button 82a or the second scroll button 52b, the document analysis support device D can display a new synonymous word display area 81, and can accept new input of co-occurrence rule words and their corresponding synonymous words through the input unit 2. By operating the confirm button 86, these can be stored in the co-occurrence rule storage unit 52. That is, multiple pairs of co-occurrence rule words and their corresponding synonymous words can be input and stored. When multiple pairs of co-occurrence rule words and their corresponding synonymous words have already been input, the document analysis support device D can sequentially display each pair in the synonymous word display area 81 by operating either the first or second scroll button 82a or 82b, and can modify the co-occurrence rule words and their corresponding synonymous words by overwriting them.

[0112] For example, a second user (second operator) different from the first user causes the synonymous word processing screen SA to be displayed on the display unit 3, and inputs "jb(junction box);terminal;number(quantity)(quantities)" using the input unit 2. When this input is accepted, "jb(junction box);terminal;number(quantity)(quantities)" as shown in FIG. 10B is displayed in the synonymous word display area 81 of the synonymous word processing screen SA. In this situation, when the second user inputs the confirm button 86 and then the end button 88, the document analysis support device D causes the synonymous word registration unit 16 to store "jb(junction box);terminal;number(quantity)(quantities)" that was input using the input unit 2 and is displayed in the synonymous word display area 81 in the co-occurrence rule storage unit 52, and then terminates the synonymous word processing screen SA.

[0113] For example, a third user (third operator) different from the first and second users causes the synonymous word processing screen SA to be displayed on the display unit 3, and inputs "junction box (JB); Ex-d (Ex d) (Ex d)" by using the input unit 2. When this input is accepted, "junction box (JB); Ex-d (Ex d) (Ex d)" shown in FIG. 10C is displayed in the synonymous word display area 81 of the synonymous word processing screen SA. In this situation, when the third user inputs the confirm button 86 and then the end button 88, the document analysis support device D causes the synonymous word registration unit 16 to store "junction box (JB); Ex-d (Ex d) (Ex d)" inputted in the input unit 2 and displayed in the synonymous word display area 81 in the co-occurrence rule storage unit 52, and then terminates the synonymous word processing screen SA.

[0114] For example, a fourth user (fourth operator) different from the first to third users causes the synonymous word processing screen SA to be displayed on the display unit 3, and then inputs "smart;I / O cabinet (junction box) (JB)" using the input unit 2. When this input is accepted, "smart;I / O cabinet (junction box) (JB)" shown in FIG. 10D is displayed in the synonymous word display area 81 of the synonymous word processing screen SA. In this situation, when the second user operates the confirm button 86 and then the end button 88, the document analysis support device D causes the synonymous word registration unit 16 to store "smart;I / O cabinet (junction box) (JB)" that was input using the input unit 2 and is displayed in the synonymous word display area 81 in the co-occurrence rule storage unit 52, and then terminates the synonymous word processing screen SA.

[0115] For example, a fifth user (fifth operator) different from the first to fourth users causes the synonymous word processing screen SA to be displayed on the display unit 3, and then inputs "smart;I / O cabinet(junction box)(jb)" using the input unit 2. When this input is accepted, "smart;I / O cabinet(junction box)(jb)" shown in FIG. 10E is displayed in the synonymous word display area 81 of the synonymous word processing screen SA. In this situation, when the second user operates the confirm button 86 and then the end button 88, the document analysis support device D causes the synonymous word registration unit 16 to store "smart;I / O cabinet(junction box)(jb)" that was input using the input unit 2 and is displayed in the synonymous word display area 81 in the co-occurrence rule storage unit 52, and then terminates the synonymous word processing screen SA.

[0116] In this situation, when a command to display a graph is received at the input unit 2, the document analysis support device D causes the display processing unit 13 to display a synonymous word processing screen SA on the display unit 3. When one of the first to fifth users described above, or a sixth user (sixth operator) different from the first to fifth users (here, the first user) performs an input operation on the graph display button 87, for example, the document analysis support device D causes the common word extraction unit 17 of the control processing unit 1 to extract words common to a plurality of co-occurrence rules and words considered to be synonymous, if the co-occurrence rule storage unit 52 stores multiple co-occurrence rules. In this third modified embodiment, the common word extraction unit 17 further extracts at least one of the words and words considered to be synonymous that are associated with the common words and words considered to be synonymous extracted by the common word extraction unit 17 as described above. For example, the common word extraction unit 17 extracts words common to multiple co-occurrence rules and words deemed to be synonyms for each co-occurrence rule registered in each record, starting from the first record stored in the co-occurrence rule storage unit 52, and further extracts at least one of the words and words deemed to be synonyms associated with the common words and words deemed to be synonyms extracted by the common word extraction unit 17. For example, in the example shown in FIGS. 10A to 10E , "jb (junction box)" input by the first and second users as described above is extracted. Here, as will be described later, if the word edit distance is the Levenshtein distance, uppercase letters are converted to lowercase letters, and the Levenshtein distance is 0 without distinguishing between uppercase and lowercase letters. Therefore, "junction box (JB)" input by the third user as described above is also extracted. Then, the "I / O cabinet (junction box) (JB)" entered by the fourth user, which is associated with the "junction box (JB)" extracted in this way, is extracted, and the "I / O cabinet (junction box) (jb)" entered by the fifth user is extracted.

[0117] After the extraction, the document analysis support device D causes the display processing unit 13 to display the common words and synonymous words extracted by the common word extraction unit 17 in a graph in the graph display area 84 of the synonymous word processing screen SA. In the example shown in Figures 10A to 10E, "jb", "junction box", "I / O cabinet", and "JB" are represented as nodes, and the graph shown in Figure 10F, in which the associations between these are represented as edges, is displayed in the graph display area 84. For example, "junction" and "JB" are associated by "junction(JB)" input by the third and fourth users as described above, and therefore the "junction" node and the "JB" node are connected by two edges.

[0118] In addition, if multiple pairs of the common words and synonymous words are extracted by the common word extraction unit 17, each pair can be displayed in sequence in the graph display area 84 by operating either the third or fourth scroll button 85a, 85b.

[0119] 9, the synonymous word processing screen SA is provided with the graph display word display area 83 as described above, and the first user may input the word for which a graph is to be displayed and the synonymous word from the input unit 2. For example, in the above example, when "junction (JB)" is input from the input unit 2 and displayed in the graph display word display area 83, and the graph display button 87 is operated, the common word extraction unit 17 executes extraction based on "junction (JB)", and the display processing unit 13 displays the graph shown in FIG. 10F in the graph display area 84.

[0120] In the document analysis support device D, document analysis support method, and document analysis support program according to the third modified embodiment, when the co-occurrence rule storage unit 52 stores multiple co-occurrence rules, the device extracts words common to the multiple co-occurrence rules and synonymous words and displays them on the display unit 3. This allows efficient revision of the ontology by referring to these words. When synonymous words are input from multiple users, synonymous words input by other users are also extracted and displayed, allowing reference to the synonymous words input by other users. For this reason, the co-occurrence rule input screen SC preferably further includes a synonymous word processing button (not shown) for inputting an instruction to display the synonymous word processing screen SA. When the synonymous word processing button is pressed, the synonymous word processing screen SA is displayed in a new window. Alternatively, the co-occurrence rule input screen SC and the synonymous word processing screen SA are preferably integrated. This allows efficient revision of the ontology by referring to the synonymous words.

[0121] The document analysis support device D, document analysis support device method, and document analysis support program in the third modified form display the common words and synonymous words extracted by the common word extraction unit 17 in graph form on the display unit 3, so that the relationship between words and synonymous words can be visually grasped and efficiently recognized.

[0122] Furthermore, in the third modified embodiment, the common word extraction unit 17 may further extract, as the common words and synonymous words, words whose word edit distance between multiple co-occurrence rules is equal to or less than a threshold (fourth modified embodiment). The threshold is set appropriately in advance. The word edit distance may be, for example, the Levenshtein distance, which is calculated as the minimum number of steps required to transform one character string into the other character string by inserting, deleting, or substituting one character between two character strings. For example, the Levenshtein distance between "kitten" and "sitting" is 3, since three steps are required: "kitten" → "sitten" → "sittin" → "sitting." Note that the word edit distance is not limited to the Levenshtein distance, and for example, the Jaro-Winkler distance may be used as the word edit distance.

[0123] Fig. 11 is a diagram for explaining, as an example, graphs in the third and fourth modified embodiments, Fig. 11A shows a graph in the fourth modified embodiment, and Fig. 11B and Fig. 11C show the graphs in the third modified embodiment.

[0124] For example, it is assumed that the synonymous word processing screen SA is displayed on the display unit 3, a first user inputs "anti surge (anti_surge)" through the input unit 2, presses the confirm button 86, and the result is stored in the co-occurrence rule storage unit 52. It is also assumed that the synonymous word processing screen SA is displayed on the display unit 3, a second user inputs "antisurge (anti-surge)" through the input unit 2, presses the confirm button 86, and the result is stored in the co-occurrence rule storage unit 52. In such a case, if "anti surge (anti_surge)" is extracted by the common word extraction unit 17 in the third modified example, "antisurge (anti-surge)" is not extracted. Therefore, the graph display area 84 of the synonymous word processing screen SA displays the graph shown in FIG. 11B, in which "anti surge" and "anti_surge" are represented as nodes and their associations are represented as edges. Similarly, in the above-mentioned third modified example, when "antisurge (anti-surge)" is extracted by the common word extraction unit 17, "anti surge (anti_surge)" is not extracted. For this reason, the graph display area 84 of the synonymous word processing screen SA displays the graph shown in FIG. 11C , in which "antisurge" and "anti-surge" are represented as nodes and their associations are represented as edges. On the other hand, in this fourth modified example, when the threshold is set to 2, for example, the Levenshtein distance of "antisurge" to "antisurge" is 1, so "antisurge" is extracted; the Levenshtein distance of "anti-surge" to "anti surge" is 1, so "anti-surge" is extracted; the Levenshtein distance of "anti_surge" to "antisurge" is 1, so "antisurge" is extracted; the Levenshtein distance of "anti-surge" to "anti_surge" is 1, so "antisurge" is extracted; and the Levenshtein distance of "anti-surge" to "anti_surge" is 1, so "anti-surge" is extracted.Therefore, the graph display area 84 of the synonymous word processing screen SA displays the graph shown in FIG. 11A, in which "anti surge", "anti_surge", "antisurge" and "anti-surge" are represented as nodes and their associations are represented as edges.

[0125] The document analysis support device D, document analysis support device method, and document analysis support program in this fourth variant extract words and synonymous words whose word edit distance is less than a threshold value as the common words and synonymous words, and therefore, as can be seen by comparing Figure 11A with Figure 11B and Figure 11C, for example, it can help a user to notice synonymous words that the same user would not have noticed.

[0126] In the fourth modified embodiment, the display processing unit 13 may further attach the word edit distance to the edge (first aspect of the fifth modified embodiment). In the example shown in FIG. 11A, a Levenshtein distance of "1" is attached to the edge connecting the "anti surge" node and the "antisurge" node ("1" is displayed near the edge). Alternatively, the display processing unit 13 may further change the display manner of the edge according to the word edit distance (second aspect of the fifth modified embodiment). The display manner of the edge is, for example, any one of the thickness of the edge, the display color of the edge, and the shade of the display color of the edge. For example, the smaller the Levenshtein distance, the thicker the edge is displayed. Alternatively, for example, the smaller the Levenshtein distance, the warmer the display color of the edge, and the greater the Levenshtein distance, the cooler the display color of the edge (for example, the display color changes sequentially from red to orange, yellow, green, light blue, blue, and purple as the Levenshtein distance increases). Alternatively, for example, the smaller the Levenshtein distance, the darker the edge is displayed.

[0127] The document analysis support device D, document analysis support device method, and document analysis support program in this fifth modified form attach a word edit distance to an edge or change the display mode of an edge depending on the word edit distance, so that the degree (strength) of association between a word and a word considered to be synonymous can be visually grasped and recognized efficiently.

[0128] In order to express the present invention, the present invention has been properly and sufficiently described above through the embodiments with reference to the drawings, but it should be recognized that those skilled in the art can easily change and / or improve the above-mentioned embodiments. Therefore, unless the changes or improvements made by those skilled in the art are at a level that causes departure from the scope of the claims described in the claims, such changes or improvements are interpreted as being included in the scope of the claims. [Explanation of symbols]

[0129] D Document analysis support device 1 Control processing section 2 Input section 3 Display section 4 Interface section (IF section) 5 Storage section 11 Control section 12 Importance determination section 13 Display processing section 14 Non-importance judgment part 15 Co-occurrence rule correction section 16 Synonymous Word Registration Section 17 Common Word Extraction Unit 51 Document storage unit 52 Co-occurrence rule memory 53 Ontology Memory 54 Important text memory section 55 Non-important text storage section 56 Misjudgment sentence storage unit

Claims

1. A display unit; a co-occurrence rule storage unit that stores one or more combined words as co-occurrence rules; an important sentence storage unit that stores one or more first sentences as important sentences; an unimportant sentence storage unit that stores one or more second sentences different from the first sentence as unimportant sentences; an ontology storage unit that stores an ontology that represents relationships between words; an importance determination unit that determines whether or not an important sentence stored in the important sentence storage unit is important based on the co-occurrence rules stored in the co-occurrence rule storage unit and the ontology stored in the ontology storage unit; a display processing unit that displays important sentences that are determined to be unimportant by the importance determination unit on the display unit; an input unit that receives input of a plurality of words as candidates for co-occurrence rules for the important sentences displayed on the display unit; an unimportance determination unit that determines whether or not the unimportant sentences stored in the unimportant sentence storage unit are important based on the candidate co-occurrence rules received by the input unit; the display processing unit displays on the display unit the candidates for co-occurrence rules received by the input unit and the determination result of importance determined by the unimportance determination unit; the display processing unit displays, on the display unit, relationships between a plurality of words as candidates for co-occurrence rules received by the input unit, based on the ontology stored in the ontology storage unit. Document analysis support device.

2. The determination result includes the number of unimportant sentences determined to be important by the unimportance determination unit. The document analysis support device according to claim 1.

3. the display processing unit displays an operation button on the display unit for displaying the unimportant sentence determined to be important by the unimportance determination unit on the display unit, and when the operation button is operated, displays the unimportant sentence determined to be important by the unimportance determination unit on the display unit.

3. The document analysis support device according to claim 1.

4. the input unit accepts changes to a plurality of words as candidates for co-occurrence rules that have already been accepted by the input unit; the unimportance determination unit re-determines whether the unimportant sentences stored in the unimportant sentence storage unit are important or not based on the candidate co-occurrence rules after the change received by the input unit; the display processing unit displays on the display unit the candidates for co-occurrence rules after the change received by the input unit and the re-determination result determined to be important by the unimportance determination unit.

3. The document analysis support device according to claim 1.

5. a co-occurrence rule correction unit that corrects the co-occurrence rules stored in the co-occurrence rule storage unit in accordance with related words that are related to the important sentences stored in the important sentence storage unit; the importance determination unit determines whether or not the important sentences stored in the important sentence storage unit are important based on the co-occurrence rules corrected by the co-occurrence rule correction unit.

3. The document analysis support device according to claim 1.

6. The related words are further related to the unimportant sentences stored in the unimportant sentence storage unit, the co-occurrence rule correction unit corrects the co-occurrence rule candidates received by the input unit in accordance with the related words; the unimportance determination unit determines whether the unimportant sentence stored in the unimportant sentence storage unit is important based on the candidate co-occurrence rule corrected by the co-occurrence rule correction unit.

6. The document analysis support device according to claim 5.

7. A document analysis support method executed by a document analysis support device including a display unit, a co-occurrence rule storage unit that stores one or more combined words as co-occurrence rules, an important sentence storage unit that stores one or more first sentences as important sentences, an unimportant sentence storage unit that stores one or more second sentences different from the first sentences as unimportant sentences, and an ontology storage unit that stores an ontology that represents a relationship between words, an importance determination step of determining whether or not an important sentence stored in the important sentence storage unit is important based on the co-occurrence rules stored in the co-occurrence rule storage unit and the ontology stored in the ontology storage unit; a display processing step of displaying important sentences determined to be unimportant in the importance determination step on the display unit; an input step of receiving input of a plurality of words as candidates for co-occurrence rules for the important sentences displayed on the display unit; an unimportance determination step of determining whether or not the unimportant sentences stored in the unimportant sentence storage unit are important based on the candidate co-occurrence rules received in the input step, the display processing step displays on the display unit the candidates for co-occurrence rules received in the input step and the determination result that the unimportance determination step determines to be important; the display processing step displays on the display unit, on the basis of the ontology stored in the ontology storage unit, relationships between the words as candidates for the co-occurrence rule received in the input step. Document analysis assistance method.

8. Computer, a co-occurrence rule storage unit that stores one or more combined words as co-occurrence rules; an important sentence storage unit that stores one or more first sentences as important sentences; an unimportant sentence storage unit that stores one or more second sentences different from the first sentence as unimportant sentences; an ontology storage unit that stores an ontology that represents relationships between words; an importance determination unit that determines whether or not an important sentence stored in the important sentence storage unit is important based on the co-occurrence rules stored in the co-occurrence rule storage unit and the ontology stored in the ontology storage unit; a display processing unit that displays on a display unit important sentences that are determined to be unimportant by the importance determination unit; an input unit that receives input of a plurality of words as candidates for co-occurrence rules for the important sentences displayed on the display unit; a text analysis support program for causing the text analysis support program to function as an unimportance determination unit that determines whether a non-important text stored in the non-important text storage unit is important based on a candidate co-occurrence rule received by the input unit, the display processing unit displays on the display unit the candidates for co-occurrence rules received by the input unit and the determination result of importance determined by the unimportance determination unit; the display processing unit displays, on the display unit, relationships between a plurality of words as candidates for co-occurrence rules received by the input unit, based on the ontology stored in the ontology storage unit. Document analysis support program.

9. A display unit; a co-occurrence rule storage unit that stores one or more combined words as co-occurrence rules; an unimportant sentence storage unit that stores one or more predetermined sentences as unimportant sentences; an ontology storage unit that stores an ontology that represents relationships between words; a target sentence acquisition unit that acquires a target sentence to be analyzed; an importance determination unit that determines whether the target sentence acquired by the target sentence acquisition unit is important based on the co-occurrence rules stored in the co-occurrence rule storage unit and the ontology stored in the ontology storage unit; a display processing unit that displays the target sentence on the display unit when the importance determination unit determines that the target sentence is not important; an input unit that receives input of a plurality of words as candidates for co-occurrence rules for the target sentence displayed on the display unit; an unimportance determination unit that determines whether or not the unimportant sentences stored in the unimportant sentence storage unit are important based on the candidate co-occurrence rules received by the input unit; the display processing unit displays on the display unit the candidates for co-occurrence rules received by the input unit and the determination result of importance determined by the unimportance determination unit; the display processing unit displays, on the display unit, relationships between a plurality of words as candidates for co-occurrence rules received by the input unit, based on the ontology stored in the ontology storage unit. Document analysis support device.

10. A display unit, a co-occurrence rule storage unit that stores one or more combined words as co-occurrence rules; an important sentence storage unit that stores one or more first sentences as important sentences; an unimportant sentence storage unit that stores one or more second sentences different from the first sentence as unimportant sentences; an ontology storage unit that stores an ontology that represents relationships between words; an importance determination unit that determines whether or not an important sentence stored in the important sentence storage unit is important based on the co-occurrence rules stored in the co-occurrence rule storage unit and the ontology stored in the ontology storage unit; a display processing unit that displays important sentences that are determined to be unimportant by the importance determination unit on the display unit; an input unit that receives input of a plurality of words as candidates for co-occurrence rules for the important sentences displayed on the display unit; an unimportance determination unit that determines whether or not the unimportant sentences stored in the unimportant sentence storage unit are important based on the candidate co-occurrence rules received by the input unit; the display processing unit displays on the display unit the candidates for co-occurrence rules received by the input unit and the determination result of importance determined by the unimportance determination unit; the input unit further receives input of synonymous words that are to be regarded as synonyms for the words of the co-occurrence rule; a synonymous word registration unit that stores the synonymous words received by the input unit in the co-occurrence rule storage unit in association with the words; a common word extraction unit that extracts words common to the plurality of co-occurrence rules and words regarded as synonyms when the co-occurrence rule storage unit stores a plurality of co-occurrence rules; The display processing unit further displays the common words and synonymous words extracted by the common word extraction unit on the display unit. Document analysis support device.

11. The common word extraction unit further extracts, as the common words and synonymous words, words whose word edit distances are equal to or less than a threshold between the plurality of co-occurrence rules. The document analysis support device according to claim 10.

12. the display processing unit displays the common words and synonymous words extracted by the common word extraction unit on the display unit in a graph in which the words and the synonymous words are represented by nodes and associations between the words and the synonymous words are represented by edges. The document analysis support device according to claim 10.

13. the common word extraction unit further extracts, as the common words and synonymous words, words whose word edit distance is equal to or less than a threshold between the plurality of co-occurrence rules; The display processing unit further attaches the word edit distance to the edge, or changes a display mode of the edge according to the word edit distance. The document analysis support device according to claim 12.

Images