JP4889964B2 - Rule writing device - Google Patents

Description

  The present invention relates to a rule text creating apparatus that can centrally manage in-house rules and prescribed requirements, and at the same time create or manage specifications and product standards.

  For example, the following Patent Document 1 discloses that in an information retrieval apparatus for retrieving a database, an input retrieval sentence in a natural language is a syntactic analysis unit that identifies a sentence structure using a term dictionary and a grammar table, and a natural language is a database. Select a correspondence table that is linked to the search key of, and identifies database items together with particles, and a correspondence table that is determined by verb correspondence from the syntax analysis result, or a correspondence table that contains many words in the table in the table. 2 shows a correspondence table using database search apparatus provided with a collation unit for creating a search logical expression and a search execution unit for collecting data based on the logical expression for database search. In this correspondence table using database search device, a natural language is converted into a search key of a database by a term dictionary and a correspondence table, a database item to be searched by a conversion process of a collation unit is designated, and an operation designated by the correspondence table It is converted to a logical expression and searched. Incorporating surface layer constraints into the correspondence table, the collation unit selects a correspondence table that contains more words in the syntax or the result of the parsing result and the correspondence table, so it is natural even for databases with few term dictionaries and hierarchical constraints Search with simple Japanese sentences. In addition, there is an effect that can easily follow the addition and change of the database and the term dictionary.

  Patent Document 2 listed below is an input means for recognizing and outputting an input database search sentence in a database search method for searching a database by creating a database command corresponding to the input sentence, and output from the input means. Analytical means for converting a database search sentence into a semantic structure, a dictionary that stores a semantic structure and a database search language in association with each other, a knowledge base that stores a database schema, and a correspondence relationship between search items and conditions on the database, A parallel relationship estimation rule describing the correspondence between the semantic structure and the parallel relationship and a relation determination means for determining the correspondence relationship on the database from the search target and condition extracted from the semantic structure with reference to the knowledge base, and a parallel relationship estimation rule Referring to, parallel estimation means for estimating parallel relations from judgment results and semantic structures, and parallel from semantic structures Database search provided with a command generation means for generating a database command for which parallel conditions are estimated using a clerk, and a database search means for searching the database using the command generated by the command generation means and outputting the result Indicates the method. In this database search method, it is possible to increase the accuracy of estimating the relationship between parallel conditions using the parallel particle “to”, and as a result, a search result close to what the searcher intends can be obtained.

Japanese Patent No. 2596260 Japanese Patent Publication No. 7-54527

However, conventionally, there has not been an apparatus that can create and manage in-house rules and rules that define prescribed requirements.
SUMMARY OF THE INVENTION An object of the present invention is to provide a rule text creation device capable of centrally managing in-house rules and prescribed requirements, and simultaneously creating or managing specifications and product standards. It is another object of the present invention to provide a rule sentence creation device capable of maintaining consistency and ensuring listability when systematically constructing, maintaining and documenting rules of an organization management system.

  The rule sentence creating apparatus according to the present invention includes a storage means (1), an input means (2) connected to the storage means (1), a calculation means (3) connected to the storage means (1), and a calculation means. Display means for displaying the output of (3). The storage means (1) is at least one or more of laws, ordinances, ministerial ordinances, ordinances, notifications, Japanese Industrial Standards, International Standards, American Society of Mechanical Engineers Standards, American Material Testing Association Standards, Arrangements and Requirements Sentences representing selected rules, multiple subjects that constitute sentences, multiple objects including direct objects and indirect objects, multiple predicates of verbs including inflectional endings, and each element of the sentence The rule database (15) for storing the protocol to be connected as storage information and the keyword information input from the input device (2) are received, and the rule database (15) is accessed based on the keyword information to store the stored information. And a search device (14) for performing a search and outputting a search result. The input means (2) includes at least one of a subject including an indirect object related to a subject and a person, a conditional clause, an object including an indirect object and a direct object related to an object, and a verb predicate including an inflection ending. Are entered as keyword information. The calculation means (3) receives the search result of the search device (14) of the storage means (1), connects the subject, object and predicate specified by the protocol to each other, and converts the search result into plaintext (natural A language processing device (34) for outputting a Japanese sentence of a rule sentence created and assembled into a plain text. The display means (4) displays the Japanese text output from the language processing device (34) of the computing means (3).

  It is possible to create or manage specifications and product standards at the same time as managing in-house rules and regulatory requirements in a centralized manner.

  An embodiment of a rule text creating apparatus according to the present invention will be described with reference to FIGS.

  As shown in FIG. 1, the rule sentence creation device of the present invention inputs information to a storage device (storage means) (1) and the storage device (1) and gives a command signal to be read to the storage device (1). Input device (input means) (2), an arithmetic device (arithmetic means) (3) for processing information read from the storage device (1), and a display device (3) for displaying the output of the arithmetic device (3) Display means) (4). The storage device (1) includes a sentence representing a rule, a plurality of subjects constituting a sentence, a plurality of objects including a direct object and an indirect object, a plurality of predicates of a verb including an inflection ending, and a sentence. A rule database (15) that stores as a storage information a protocol that connects the elements to each other, and a search device (14) that accesses and searches the storage information of the rule database (15) based on the keyword information. The element constituting the sentence may include a conditional clause. The rules that make up the text are selected from at least one or more of laws, ordinances, ministerial ordinances, rules, ordinances, Japanese Industrial Standards, International Standards, American Society of Mechanical Engineers Standards, American Material Testing Association Standards, Arrangements Is done. The elements for which a rule is satisfied are: rule identification number, rule title (classification, proper noun), subject that executes the rule (subject), rule that requires the rule to be enforced (predecessor rule), and conditions required for the rule ( Trigger), when and how often the rule is implemented, the procedure (protocol) for implementing the rule, the results of the rule implementation, who should report the results of the rule implementation (indirect object), should be attached after the rule implementation Includes rules (successor rules), predicate verbs when implementing the rules, superordinate rules of rules, subordinate rules of rules.

  As shown in FIG. 1, the input device (2) has a keyboard (21) normally used in an electronic computer and a display (22) connected to the keyboard (21). Although not shown, the input device (2) includes a CPU having known arithmetic functions such as character conversion and code conversion of an input signal from the keyboard (21). The information from the input device (2) is converted into a corresponding address signal through the input language processing device (31) of the arithmetic device (3) and given to the storage device (1). Storage device as keyword information of at least one of a subject including an indirect object related to a subject and a person, a conditional clause, an object including an indirect object related to an object and a direct object, and a verb predicate including an inflected ending It is possible to input from the input device (2) to (1).

  The storage device (1) receives a signal from the input language processing device (31) of the arithmetic device (3), analyzes the meaning of the signal, and distributes the output, and a semantic analysis device A dictionary maintenance device for storing and retrieving a plurality of predicates including a plurality of subjects including a plurality of subjects, direct objects and indirect objects, and predicate verbs and endings as elements constituting the sentence from (11) ( 12) and a storage execution device (13) for storing rules constituting the sentences from the semantic analysis device (11) and rules constituting the sentences and protocols for connecting each element of the sentences in the rule database (15) And a search device (14) for searching for information stored in the rule database (15) according to the address signal received from the semantic analysis device (11) and outputting the searched information to the arithmetic device (3). Yes.

  The arithmetic device (3) receives the signal from the rule database (15) of the storage device (1), processes the meaning of the signal received by the receiver (32), and generates an output. A semantic processing device (33), and a language processing device (34) that composes an output from the semantic processing device (33) into a sentence. The language processing device (34) connects the subject, the object and the predicate specified by the protocol read from the storage means (1) to each other, and outputs the result composed of sentences. The display device (4) can receive the output from the language processing device (34) of the arithmetic device (3), display the output on the display (41), or print it by the printer (42).

  The storage device (1) of the rule sentence creation device stores rules constituting the sentence and a protocol for interconnecting each element of the sentence according to the procedure shown in FIG. In step 50 of FIG. 2, when a rule sentence or protocol is input from the input device (2), it is sent to the semantic analysis device (11) through the input language processing device (31), and the semantic analysis device (11) is the same. It is determined whether or not the information already exists in the electronic dictionary (16) (step 52). As shown in FIG. 4, “subject” including “indirect object” related to “subject” and “person”, “conditional clause”, “indirect object” related to thing, and “direct object” including “direct object”, “verb” All the rules are entered into the input device (2) separately for each element including the “end” of the “utilized” type, and all rules are connected by connecting multiple rule sentences to each other by the “protocol”. Centralize sentences and define set rule sentences as related aggregates. For example, as shown in FIG. 5, by assigning a related set of set rule sentences and unique attributes to the target, the rule simple sentences that are all the targets (including “concept”) are associated and stored in a general format. be able to. In the present embodiment, the subject, the indirect object, and the direct object are, for example, as phrases including an appendage or a particle attached to a word, such as “to”, “to”, and “to”, respectively. Stored in the rule database (15). Similarly, verbs and endings are stored in the rule database (15) together with stems and inflection endings such as “manage” and “check”. Usually, in the arithmetic unit (3), a program that creates a sentence including a subject and a predicate following the subject is executed, but a program that creates an abbreviated sentence without a subject, an inverted sentence in which the subject and the predicate are reversed. Can be executed.

  If the same information does not exist in the electronic dictionary (16), the semantic analyzer (11) proceeds to step 53 and determines whether or not to add to the electronic dictionary (16). At this time, if the operator instructs not to add by pressing the non-add key of the input device (2), the process proceeds to step 57, and the process is terminated. In the case of adding in step 53, when a specific key is pressed to instruct addition, it is transmitted to the dictionary maintenance device (12), and an inquiry is made as to whether the "meaning of the rule" is correct (step 55). . This inquiry is displayed from the dictionary maintenance device (12) on the display (22) of the input device (2) via the semantic analysis device (11) and the input language processing device (31). At this time, when the operator presses a specific key of the input device (2) to instruct that the meaning is correct, the dictionary maintenance device (12) causes the electronic dictionary (16) to store and register the rule ( Step 58). When the “meaning of the rule” is incorrect in step 55, the process proceeds to step 59, the meaning of the rule is changed, input from the input device (2), and the process returns to step 54. If the “meaning of the rule” is correct in step 55, the process proceeds to step 58 to store and register the rule.

  In step 52, if a rule already exists in the electronic dictionary, the process proceeds to step 56, and an inquiry is made as to whether or not to change the rule. If the rule is not changed, the process proceeds to step 57 and the process is terminated. When changing the rule, after the operator presses a specific key of the input device (2), the changed rule is input and the process returns to step 52, and the procedure after step 52 is performed.

  Next, when carrying out the rule text creation method, if keyword information is input from the input device (2), the input keyword information is sent to the search device (14) of the storage device (1), and the search device ( 14) accesses the rule database (15) based on the keyword information to search the stored information, sends the search result to the language processing device (34) of the arithmetic device (3), and the language processing device (34 ) Connects the subject, object and predicate specified by the protocol to each other, composes the search results into plain text (natural text), assembles them, and outputs the Japanese output of the rules created in plain text to the display device (4) Output. That is, as shown in step 61 of FIG. 3, when a keyword in which a part of the rule sentence indicates “who / what?” Is input from the input device (2), the input keyword information is converted into an input language processing device ( It is given to the semantic analyzer (11) via 31). The semantic analysis device (11) recognizes the keyword that is the search term (step 62) and sends it to the search device (14) (step 63). The search device (14) accesses the rule database (15). The search is performed (step 64), and the search result is sent to the semantic processing device (33) and the language processing device (34) through the input device (32) of the arithmetic device (3) (step 65). The language processing device (34) composes the search results in plain text (natural text) and assembles them (step 66). Thereafter, the language processing device (34) generates the output of the Japanese text of the rule created in plain text on the display device (4) (step 67), and the Japanese text is displayed on the display (41) or the printer ( 42) is printed.

  When creating a rule text for each department, first, a “subject” is input to the input device (2), and rules having the designated “subject” can be listed. In this state, a specific “operation” or / and “object” can be input to the input device (2) to list all the activities to be performed by the specific organizational unit and output the division of work. .

  When creating a rule sentence according to the business flow, enter the `` object '' and `` indirect object '' related to `` person '' into the input device (2), the entered `` indirect object '' as the subject, and `` object '' Can be extracted as a “direct object”. Similarly, a series of rules that are semantically continuous can be extracted hierarchically (recursively). In addition, it shows the relevance of the business (process) and outputs it. As a result, a series of business flows can be clarified, and continuity and consistency can be verified.

  Also, it is possible to input subject, object, or subject, verb, etc., identify all similar tasks in the same department, extract duplicate rules and approximate rules, and organize and consolidate the tasks.

  The present invention provides an apparatus and method that can easily create and manage in-house rules and rules that define prescribed requirements.

FIG. 1 is a block diagram schematically showing an apparatus for creating rule text according to the present invention. The flowchart which shows the procedure which divides | segments a single sentence rule for every element, and inputs into an input device, when implementing the rule text preparation method by this invention The flowchart which shows the procedure which produces a rule sentence using the rule sentence preparation apparatus by this invention Flowchart showing the structure of a structured rule and the procedure for entering the rule database Block diagram showing related aggregate attributes

Explanation of symbols

  (1) ・ ・ Storage device (storage means), (2) ・ ・ Input device (input means), (3) ・ ・ Calculation device (calculation means), (4) ・ ・ Display device (display means), (11・ ・ Semantic analysis device, (12) ・ ・ Dictionary maintenance device, (13) ・ ・ Storage execution device, (14) ・ ・ Search device, (15) ・ ・ Rule database, (16) ・ ・ Electronic dictionary, ( (21) ・ ・ Keyboard, (22) ・ ・ Display, (31) ・ ・ Input language processing device, (32) ・ ・ Receiving device, (33) ・ ・ Semantic processing device, (34) ・ ・ Language processing device, ( 41) ・ ・ Display, (42) ・ ・ Printer,

Claims (2)

A storage means, an input means connected to the storage means, a calculation means connected to the storage means, and a display means for displaying the output of the calculation means,
The storage means is selected from at least one or more of laws, government ordinances, ministerial ordinances, ordinances, Japanese Industrial Standards, International Standards, American Society of Mechanical Engineers Standards, American Material Testing Association Standards, Arrangements and Requirements Sentences that represent rules, multiple subjects that make up sentences, multiple objects that include direct objects and indirect objects, multiple predicates of verbs that contain inflection endings, and elements of sentences A rule database for storing a protocol as storage information; and a search device for receiving keyword information input from an input device, accessing the rule database based on the keyword information, searching the storage information, and outputting a search result; Have
The input means is a keyword including at least one of a subject including an indirect object related to a subject and a person, a conditional clause, an object including an indirect object and a direct object related to an object, and a verb predicate including an inflection ending. Enter as information,
The calculation means receives the search result of the search means of the storage means, connects the subject, object and predicate specified by the protocol to each other, composes the search result into plain text (natural text), assembles it into plain text It has a language processing device that outputs the Japanese text of the created rule text,
The rule text creating apparatus, wherein the display means displays Japanese text output from the language processing device of the computing means.   A sentence that represents a rule, a plurality of subjects, objects, and predicates that are elements constituting the sentence, and a protocol that interconnects each element of the sentence that represents the rule are stored in the storage means through the input means. 1. The rule sentence creation device according to 1.

Images