JP5581861B2 - Retrieval device, method and program, and data parsing device having retrieval function - Google Patents

Description

  The present invention relates to search processing technology. The present invention also relates to text data segmentation processing technology using the search processing technology.

  In the information integration field where master data is integrated, collected data is used, etc., parsing and cleansing are generally performed in order to make it easy to handle input text data (character strings).

  Parsing is a process that divides the input text data (character string) into individual elements so as to match a predetermined rule that defines the arrangement of elements, and specifies the attributes of each divided element. It is.

  Cleansing is a process of arranging input text data into a database storage format, and is a process of unifying full-width / half-width, notation, and the like. Parsing processing occupies an important position when text data that is “solid” like Japanese is handled as database data.

  As an example of text data that needs to be parsed, there is database master data. Most of these text data are customer data, and in particular, there are many data including unique names such as company names and organization names. Text data including such a unique name is corporate data. The corporate data is, for example, text data including a corporate name such as a company or organization, a text data including a name associated with the corporate name, for example, a corporate personality name indicating a corporate personality, a department name indicating a corporate internal organization, etc. Etc.

  In order to use corporate data as master data, a parsing process that identifies each of one or more names included in the corporate data and identifies the attribute of each name is essential.

  As a parsing process, a unique name is registered in the dictionary in advance, the registered words in the dictionary appearing in the corporate data to be analyzed are searched, the element sequence that is most likely to be arranged is selected, The technique of dividing into elements (names) is common.

  A typical example is morphological analysis processing. In addition, there is known a method of focusing on the items in the corporate data to be analyzed and dividing each element by combining forward matching or backward matching with dictionary registered words.

  In addition, when an element obtained by sharing corporate data is regarded as an abbreviation, there is known a method for determining what the official name corresponding to the abbreviation is.

Japanese Patent No. 4076900 Japanese Patent No. 2792147 JP-A-4-130578 JP-A-11-353316

  Generally, the name registered in the parse processing dictionary is the official name. However, since input data to be processed is manually input and written, it is often expressed by an abbreviated name rather than an official name due to labor saving. Therefore, it is required to identify the official name from the abbreviations that appear in the input data.

  Therefore, it is conceivable that the abbreviations that are expected to appear and the correspondence between the abbreviations and the official names are registered in the dictionary as much as possible. However, the abbreviations are expected to be enormous, and it is difficult to cover the abbreviations, which is not a practical method.

  Specifically, in the case of the conventional method of preparing a dictionary for registering the proper names of companies, organizations, etc. as parsing processing dictionaries, and searching the registered words (official names) of the dictionary from the input data. Because only registered words that completely match the input data to be searched can be searched, the following problems exist.

  Problem 1: Even if the official name is registered in the dictionary, the official name expressed by the abbreviation in the input data cannot be determined.

  A method for dealing with the problem 1 is to “register all abbreviations that may be used in the input data in the dictionary”. However, it is practically difficult to prepare data that covers abbreviations in a uniform number.

  Problem 2: As a related name, a name derived from a parent name (hereinafter referred to as a derived name) is often used. For example, as a corporate name of a newly established subsidiary, a derivative name of the form “parent company name + α” is often named. In this case, if only the official name of the parent company is registered in the dictionary, the name of the subsidiary (derived name) included in the input data will be divided into “name of the parent company” and “α”. There is a problem that it is not identified as a name.

  As with the abbreviation, the coping method for the problem 2 is “Register all related names in the dictionary”. However, since there are not only a large number of related names but also frequent updates, it is practically difficult to prepare a comprehensive dictionary.

  Furthermore, as a method of associating the registered words (official names) in the dictionary with the abbreviations in the input data as a conventional method, “Registered words in which all characters appearing in a string that can be regarded as abbreviations in the dictionary appear in that order.” If there is, there is a process of “determining the registered word as an official name for the abbreviation”.

  This conventional method can associate an abbreviation with a formal name. However, as a premise, it is necessary to cut out “a character string that can be regarded as an abbreviation” from the input data by some method, but a method for automatically cutting out the range of the abbreviation from the input data is not known.

  As another conventional method, there is a method in which a pattern dictionary in which a pattern is registered instead of the name itself is prepared, and the input data is divided into each name by matching the pattern dictionary with the input data. For example, it is determined that “a character string ending with“ part ”is“ department name ””, and the range of the department name is specified and divided.

  However, in pattern matching processing, it is necessary to prepare a dictionary that indicates a pattern that may be applicable, not the name itself, and it is difficult not only to describe element patterns according to certain rules. , It is difficult to describe an accurate pattern that can correspond to the elements to be identified, and there is a problem of imposing a heavy burden on the user.

  An object of the present invention is to provide an apparatus capable of performing a process of specifying abbreviations and derived names included in input data to be processed and a process of searching for the formal names with higher accuracy using a dictionary in which only formal names are registered. It is to be.

  Another object of the present invention is to provide the above-described search processing function, divide a portion regarded as an abbreviation or a derived name in input data from other portions, and form an official name corresponding to the identified abbreviation or derived name. It is an object of the present invention to provide a device capable of executing a parsing process for associating a.

  Furthermore, the above and other objects and novel features of the present invention will become apparent from the description of the specification and the accompanying drawings.

A search device disclosed as one embodiment of the present invention includes 1) a storage unit that stores a plurality of registered formal names , 2) a data input unit that receives an input character string, and 3) a processing unit. When the input character string does not match any of a plurality of registered formal names , the processing unit stores a plurality of character strings including character strings partially matching the input character string from the storage unit . retrieved from official name, among the search results extracted official name, a length of the input string and partially string that matches the maximum and partial match to the length of the official name A formal name that maximizes the ratio of the length of the character string to be identified is identified, and a partially matching character string is determined as a character string indicating an abbreviation corresponding to the identified formal name .

The search method disclosed as another embodiment of the present invention is a processing method executed by a computer having a storage unit, which includes 1) a processing step for receiving an input character string, and 2) a plurality of registered methods . A character including a character string partially matching the input character string from the storage unit when the input character string does not match any of a plurality of registered formal names with reference to the storage unit storing the official name A processing step for retrieving a column from a plurality of registered formal names , and 3) of the formal names extracted as a result of the retrieval, the character string partially matching the input character string has a maximum length, and , letters indicating the abbreviations length ratio of the character string that partially matches to the length of the official name identifies a formal name of maximum, the string that partially matches, corresponding to the identified official name Processing step for determining a row It is those with a door.

Furthermore, another search programs which are to be disclosed as one embodiment of the present invention, the computer having a serial憶部is intended for executing a process implemented by the processing steps described above.

Further, disclosed as another embodiment of the present invention, the data parsing apparatus comprising a search function, a character string as a component a plurality of names, the proper arrangement order of a plurality of names as a string based on the combination rule shown, a data parsing apparatus comprising a parsing unit for executing parsing process of dividing each name that is a component, 1) you store name dictionary to register words the official name Symbol憶And 2) a data input unit that accepts an input character string , and 3) if the input character string includes a part that does not match the registered word in the name dictionary , the registered word in the name dictionary is compared with the input character string and registered A word and a partial character string that partially matches the registered word are extracted in association with each other, and among the extracted registered words, the length of the associated partial character string is the largest and the ratio to the input character string is the largest. Identify one registered word And, determining that abbreviation a substring associated with the identified registered words of the input string, and a search processing unit to the full name of abbreviated registered word and the particular, the parsing unit, A part determined as an abbreviation of the input character string in the parsing process is divided as one name, and a process of associating the specified formal name with the abbreviation is performed.

According to the disclosed search device, using the dictionary registered only formal name identifies the abbreviations contained in the input data to be processed, it is possible to search the full name that corresponds to the abbreviation .

Further, according to the data parsing apparatus having the disclosed search function, if it contains abbreviations in the input data to be processed, by using the dictionary registered only official name, a portion regarded as abbreviation A parsing result can be output by identifying and dividing as one element and associating a formal name corresponding to the abbreviated name.

It is a figure which shows the block structural example of the data parsing apparatus in one embodiment of this invention. It is a figure which shows the structural example of the dictionary memorize | stored in the dictionary memory | storage part in this embodiment. It is a figure which shows the structural example of the combination rule memorize | stored in the combination rule memory | storage part in this embodiment. It is a figure which shows the example of the corporate data which the data input part of the data parsing apparatus in this embodiment receives an input. It is a figure which shows the example of the parse result output from the result output part of the data parsing apparatus in this embodiment. It is a figure which shows the outline | summary flow of a process of the data parsing apparatus in this embodiment. It is a figure which shows the outline | summary flow of an abbreviation search process (step S5). It is a figure for demonstrating the partial matching search of a search object and a registration word. It is a figure which shows the specific example (the 1) of the determination process (step S17 and S18) in an abbreviation search process. It is a figure which shows the specific example (the 2) of the determination process (step S17 and S18) in an abbreviation search process. It is a figure which shows the specific example (the 3) of the determination process (step S17 and S18) in an abbreviation search process. It is a figure which shows the detailed flow of a search process (step S10) of partial matching. It is a figure which shows the detailed process flow of the search process (step S101) of back partial matching. It is a figure for demonstrating the search process of a back partial match. It is a figure which shows the detailed processing flow of a search process (step S103) of a front partial match. It is a figure which shows the detailed flow of a complete inclusion search process (step S11). It is a figure which shows the detailed flow of a complete inclusion search process (step S11). It is a figure which shows the parsing process example (the 1) of the corporate data input into the data parsing apparatus in this embodiment. It is a figure which shows the parsing process example (the 2) of the corporate data input into the data parsing apparatus in this embodiment. It is a figure for demonstrating the abbreviation search process which decomposes | disassembles and uses the registration word of a department name dictionary. It is a figure which shows the hardware structural example for implementing the data parsing apparatus in this embodiment.

  FIG. 1 is a diagram showing a block configuration example of a data parsing apparatus according to an embodiment of the present invention.

  The data parsing apparatus 1 is an apparatus that receives an input of corporate data, searches for a formal name corresponding to an abbreviation or derivative name included in the input corporate data, and parses the corporate data using the search result. .

  The corporate data is a character string including a unique name, and in this embodiment, a corporate name such as a company name or an organization name, a corporate case name, or a department name that is the name of an internal organization of the corporation may be included. Shall.

  The abbreviation is a name expressed by omitting a part of the official name. For example, in the case where the official name is “Fuji ○ abcdef”, the name is represented as “Fuji ○ ab ab”.

  A derived name is a name that is derived by adding a character string to the official name. For example, when the official name is “wet ○ dori”, it is a name represented as “wet bw xyz”.

  The data parsing device 1 includes a parsing control unit 10, a dictionary storage unit 11, a combination rule storage unit 12, a data input unit 13, a dictionary lookup unit 14, an abbreviation search unit 15, a parsing unit 16, and a result output unit 17.

  The parse control unit 10 controls processing of the data input unit 13, dictionary lookup unit 14, abbreviation search unit 15, parse unit 16, and result output unit 17. The parse control unit 10 processes the corporate data (hereinafter referred to as input data) received from the data input unit 13 by calling each processing unit of the dictionary lookup unit 14, the abbreviation search unit 15, and the parse unit 16, and obtains the processing result. It passes to the result output unit 17.

  The dictionary storage unit 11 is a dictionary used by the dictionary lookup unit 14 and the abbreviation search unit 15 and stores a dictionary in which formal names (hereinafter referred to as registered words) such as corporate names, corporate case names, and department names are registered. To do. Registered words are basically formal names, but may include names that are considered formal names.

  In the dictionary stored in the dictionary storage unit 11, as registered words, corporate names such as “Fuji * tsu” and “Fuji * tsu Production” and corporate names such as “corporation”, “limited company”, and “joint company” are registered. be registered. Department names such as “General Affairs Department”, “Sales Department”, “Business Department”, and “Management Department” may be registered in the dictionary storage unit 11.

  FIG. 2 is a diagram illustrating a configuration example of a dictionary stored in the dictionary storage unit 11.

  As shown in FIG. 2A, the dictionary storage unit 11 is configured separately for each registered name, such as “corporate name dictionary”, “corporate case name dictionary”, and the like. Alternatively, as shown in FIG. 2B, a registered word may be described in a format such as “name + name type”, and all the names may be registered in one dictionary. Furthermore, the dictionary storage unit 11 may include a “department name dictionary”.

  The combination rule storage unit 12 stores combination rules used in the parsing unit 16 that describe the correct order of the names.

  FIG. 3 is a diagram illustrating a configuration example of combination rules stored in the combination rule storage unit 12.

  As combination rules, “corporate case name + corporate name + department name”, “corporate name + corporate case name + department name”, and the like are registered. “Corporate name + corporate name + department name” indicates that the input data is composed of a sequence of “corporate case name”, “corporate name”, and “department name”.

  The data input unit 13 receives input of corporate data to be processed into the data parsing device 1 by a user operation or a request from another computer.

  FIG. 4 is a diagram illustrating an example of input corporate data.

  The corporate data includes, for example, “Fujitsu Day Co., Ltd. Day XX Management / Moldware Division”, “Fukumi Ltd. Professional”, “Fumitsu ○ Muware”, etc.

  The dictionary lookup unit 14 receives the input data from the parse control unit 10, searches the dictionary storage unit 11 for a registered word that completely matches the character string of the input data, and returns the search result to the parse control unit 10.

  The abbreviation search unit 15 receives the input data from the parse control unit 10, and when a portion that does not completely match the registered word in the dictionary storage unit 11 remains in the input data, the character string that is the search target of the input data Is searched for a partially registered word having a partially matching range, or a registered word that is not completely matched but completely included, and returns the search result to the parse control unit 10.

  The parsing unit 16 receives the processing result of the dictionary lookup unit 14 or the abbreviation search unit 15 from the parsing control unit 10, and receives input data so as to form an array of names defined in the combination rules stored in the combination rule storage unit 12. Are parsed, a parsing process for specifying the attribute of each name is performed, and the result is returned to the parse control unit 10.

  The result output unit 17 outputs the result of the parsing process of the input data.

  FIG. 5 is a diagram illustrating an example of a parsing result output from the result output unit 17.

  In the parsing result, the character string of the input data is divided into one or more names, and attributes of the divided names (corporate name, legal case name, department name, etc.) are specified. Further, if the divided name is an abbreviation, an abbreviated formal name is added, or if the divided name is a derived name, a related formal name is added.

  For example, if the input data is corporate data “Tofu ○ Toshi Co., Ltd. Day XX Management and Software Division”, “Corporate name = Tofu ○ Toshi, Corporate case name = Incorporated, Department name = D A parsing result such as “Nementment / Mileware Division, Corporate Name_Official Name = Wealth ○ Tong” is output.

  In addition, when the input data is corporate data “Fujitsutsu Co., Ltd.”, “Corporate name = Fujitsutsu Pro <abbreviation>”, Corporate name = corporation, Department name = None (−), Corporate name_ A parsing result such as “official name = wealth ○ through production” is output.

  Furthermore, when the input data is corporate data “Tomi * tsu Miruware”, “Corporate name = Tomi * tsumi * ware <group>, Corporate name = None (−), Department name = None (−) , Such as “corporation name_official name = wealth ○ communication” is output.

  Next, the processing flow of the data parsing apparatus 1 will be described.

  When the data input unit 13 receives input of corporate data including an abbreviation, the parsing control unit 10 of the data parsing apparatus 1 controls the parsing process for the input corporate data.

  FIG. 6 is a diagram showing an outline flow of processing of the data parsing apparatus 1.

  Step S1: The parse control unit 10 passes the corporate data (hereinafter referred to as “input data”) accepted by the data input unit 13 to the dictionary lookup unit 14. The dictionary lookup unit 14 refers to the dictionary storage unit 11, searches all registered words (corporate name, corporate case name, department name) included in the character string of the input data, and returns the search result to the parse control unit 10. To do.

  Step S2: The parse control unit 10 checks whether or not a portion that does not match the registered word remains in the input data from the search result of the dictionary lookup unit 14.

  If there is no remaining part of the input data that does not match the registered word (all the character strings of the input data match the registered word) (N in step S2), the input data is composed only of the registered official name. Therefore, the parse control unit 10 proceeds to the process of step S3 in order to execute a general parse process. On the other hand, if a portion that does not match the registered word remains in the input data (Y in step S2), there is a possibility that the input data includes a part of the abbreviation or derived name. In step 10, the process proceeds to step S4.

  Step S3: The parse control unit 10 passes the input data and the processing result of the dictionary lookup unit 14 to the parse unit 16. The parsing unit 16 performs a parsing process that divides input data into each name so as to match the combination rules of the combination rule storage unit 12. Note that the parsing process executed by the parsing unit 16 is known and will not be described in detail.

  Step S4: The parsing control unit 10 identifies the part of the legal case name in the input data, sets the part other than the legal case name in the input data as a character string (search target) to be processed by the abbreviation search unit 15, and selects the search target. Pass to the abbreviation search unit 15.

  If there is no legal name in the input data, the parse control unit 10 searches the entire input data. Further, if the character string from the beginning of the input data is the legal name, the parse control unit 10 searches the range from the character immediately after the legal name in the input data to the last character.

  In addition, when the legal case name is in a portion other than the character string from the beginning in the input data, the parse control unit 10 searches the range from the beginning of the input data to the character immediately before the legal case name as a search target for the legal name. The range from the character immediately after the legal name to the last character is searched for the department name. In this case, the abbreviation search unit 15 searches for the abbreviations of the corporate name and the department name for the two search targets, respectively.

  When the registered word in the department name dictionary is searched from the input data in the dictionary lookup section 14, the parse control section 10 searches the range other than the legal case name and the department name from the input data as an abbreviation search section 15 for searching. May be passed to

  Step S5: The abbreviation search unit 15 performs abbreviation search processing for the search target. Details of the abbreviation search process will be described later.

  Step S6: When a search result is obtained by the abbreviation search process (Y of Step S6), the parse control unit 10 proceeds to the process of Step S7, and when the search result is not obtained (N of Step S6) ), The process proceeds to step S8.

  Step S <b> 7: The parse control unit 10 passes the input data, the processing result of the dictionary lookup unit 14, and the processing result of the abbreviation search unit 15 to the parse unit 16.

  In accordance with the combination rule of the combination rule storage unit 12, the parsing unit 16 assigns the input data to each name (character string) based on the processing result of the dictionary lookup unit 14 and the search result of the abbreviation search unit 15. ).

  If the legal name is not specified in the input data, the parsing unit 16 determines each element obtained by dividing the input data by a combination rule such as “corporate name” and “corporate name + department name”. Each parsing name is generated, and a parsing result with a legal name = none (−) is generated. Further, the parsing unit 16 adds the official name corresponding to the abbreviation if it is determined that the corporate name of the input data is “abbreviation” as a result of the abbreviation search, and is determined to be “derived name”. For example, add the official name related to the derived name.

  Step S8: The parse control unit 10 passes the input data and the processing result (none) of the abbreviation search unit 15 to the parse unit 16. The parsing unit 16 divides only the names that can be identified from the input data based on the combination rules and processing results of the combination rule storage unit 12.

  Step S <b> 9: The parse control unit 10 passes the processing result obtained from the parse unit 16 to the result output unit 17. The result output unit 17 outputs the passed result in a predetermined format.

  FIG. 7 is a diagram showing an outline flow of the abbreviation search process (step S5).

  Step S10: The abbreviation search unit 15 searches a part of the search target that partially matches the registered word, and sets the result as R1. Specifically, the abbreviation search unit 15 compares each registered word of the corporate name in the dictionary storage unit 11 with the search target, and determines the correspondence between the matching character string portion (matching range) and the registered word. Store in result list.

  FIG. 8 is a diagram for explaining a partial match search between a search target and a registered word.

  As shown in FIG. 8A, the abbreviation search unit 15 compares the character string from the end of each registered word of the corporate name stored in the dictionary storage unit 11 with the character string from the start of the search target. . For example, if the character string from the beginning of the search target “Tomi * tsu Pro Planning and Production Department” is compared with the character string from the end of the registered word “Toyama Tomi * tsu”, “Tomi * tsu” matches. To do. The abbreviation search unit 15 stores the correspondence between the character string matching part (matching range) “Tomi * tsu” and the registered word “Toyama Tomi * tsu” in the list of the result R1.

  Further, as shown in FIG. 8B, the abbreviation search unit 15 obtains a character string from the beginning of each registered word of the corporate name stored in the dictionary storage unit 11 and a character string from the beginning of the search target. Compare. For example, if the character string from the beginning of the search target “Tomi * tsu Pro Planning and Production Department” is compared with the character string from the beginning of the registered word “Tomi * tsu Production,” “Tomi * tsu Pro” matches. . The abbreviation search unit 15 stores in the result list the correspondence between the character string matching portion (matching range) “Fujitsutsu Pro” and the registered word “Fuji ○ tsu Production”. The abbreviation search unit 15 stores the processing result of all registered words of the corporate name stored in the dictionary storage unit 11 and the search target in the result list, and sets the result with the longest matching range as “R1”.

  Step S11: The abbreviation search unit 15 searches for a part that completely includes the registered word in the search target, and sets the result as R2.

  The abbreviation search unit 15 compares the character string from the top of each registered word of the corporate name stored in the dictionary storage unit 11 with the character string from the top of the search target. For example, if the character string from the beginning of the search target “Fujitsu Tsutsu Pro Planning and Production Department” is compared with the character string from the beginning of the registered word “Fujitsutsu”, Included in the search target. The abbreviation search unit 15 stores the completely included registered word “wealth *” in the result list. The abbreviation search unit 15 compares all registered words of the corporate name stored in the dictionary storage unit 11 with the search target, stores the processing result in the result list, and obtains the result with the longest matching range (the longest registered word). ) Is “R2”.

  Step S12: The abbreviation search unit 15 determines whether only R1 has been obtained as the processing result of Step S10 and Step S11. If only R1 is obtained as a processing result (Y in step S12), the process proceeds to step S13. If only R1 is not obtained (N in step S12), the process proceeds to step S14.

  Step S13: The abbreviation search unit 15 determines the partially matched character string in the search target as “abbreviation”, and sets the registered word of R1 as the abbreviated “official name”.

  Step S14: The abbreviation search unit 15 determines whether only R2 has been obtained as the processing results of steps S10 and S11. If only R2 is obtained (Y in step S14), the process proceeds to step S15. If only R2 is obtained (N in step S14), the process proceeds to step S16.

  Step S15: The abbreviation search unit 15 determines that the character string to be searched is “derived name”, and sets the registered word of R2 as “official name”.

  Step S16: The abbreviation search unit 15 determines whether both R1 and R2 are obtained as the processing results of Step S10 and Step S11. If both R1 and R2 are obtained (Y in step S16), the process proceeds to step S17. If both R1 and R2 are not obtained (N in step S16), the process proceeds to step S19. .

  Step S17: The abbreviation search unit 15 compares the character string lengths of the matching range of R1 (partial matching range) and the matching range of R2 (completely included registered word), and selects a result with a longer matching range. Or, if the matching range is the same length, select a result that has a shorter registered word length.

  Step S18: The abbreviation search unit 15 determines that the search target is “abbreviation” if the selected result is R1, and “derived name” if it is R2, and selects the registered word of the selected result as the search target. The official name.

  Step S19: The abbreviation search unit 15 outputs “none” as a result.

  9 to 11 are diagrams illustrating specific examples of the determination process (steps S17 and S18) in the partial match search process.

  As shown in FIG. 9, when the search target is “Fujitsu Production Department”, the result R1 of the partial match search is “abbreviation = wealth ○ production professional, official name = wealth ○ production production”. , Suppose that the result R2 of the complete inclusion search is “derivative name = wealth ○ communication production department, formal name = wealth ○ traffic”.

  In this case, since the length of the matching range between the search target in R1 and the registered word = 5 (Fuji * tsu Pro) and the length of the matching range at R2 = 3 (Fuku * tsu), the result of R1 Is selected. Then, as a result of the selection of R1, the result “abbreviation = wealth ○ communication, formal name = wealth ○ communication production” is returned to the parse control unit 10.

  After that, the parsing unit 16 parses the input data “Fujitsu Production Production Department” based on this result, and the result output unit 17 outputs the parsing result (No. 2) shown in FIG.

  As another example, as shown in FIG. 10, when the search target is “wealth * communication”, the result R1 of the partial match search is “abbreviation = wealth * communication, formal name = wealth * communication production”. And the result R2 of the complete inclusion search is “derivative name = wealth * communication, official name = wealth * trait”.

  In this case, since the length of the matching range between the search target in R1 and the registered word = 3 (wealth circle) is the same, the lengths of the matched registered words are compared. In other words, the proportion of matching parts among the matching registered words is calculated. In the case of R1, the number of registered character strings = 10, and in the case of R2, the number of registered character strings = 3, so the result of R2 is selected. As a result of the selection of R2, the result “Derived Name = Wide ○ Mulware, Formal Name = Wide ○ Way is returned to the parse control unit 10. As in FIG. 9, the parse result ( No. 3) is output.

  As another example, as shown in FIG. 11, when the search object is “Foo second building”, the result R1 of the partial match search is “abbreviation = Foo, official name = Studio F It is assumed that the result R2 of the complete inclusion search is “none”.

  In this case, since there is no result of R2, the result of R1 is selected, and the result “abbreviation = Foot, official name = StudioFoot” is returned to the parse control unit 10, as shown in FIG. Similarly, the parse result (No. 5) shown in FIG. 5 is output.

  As described above, even if unregistered abbreviations and derived names are included in the input data, a more appropriate parsing result can be output using registered words.

  For example, in the example shown in FIG. 9, in the case of the conventional method, it is parsed as “corporate name = wealth ○ communication, department name = professional production department”, and “we wealth ○ communication professional” is the official name (we wealth ○ communication production). Could not be identified as the corporate name of the abbreviation. Further, in the example shown in FIG. 10, in the case of the conventional method, “corporate name = wealth * communication, department name = milware” is parsed, and “fortune * communication” is changed to the official name of the parent corporation ( It was not possible to specify the name as a derivative name from Tomi * tsu).

  In the example shown in FIG. 11, if the conventional method is used, the corporate name cannot be searched, the search result = “none”, and “Footsu” is an abbreviation for the official name (Studio Footsu). It could not be specified as a corporate name. It can be seen that the data parsing apparatus 1 can identify the input data “Foot” as an abbreviated corporate name.

  Next, each processing step in the processing flow shown in FIG. 7 will be described in more detail.

  FIG. 12 is a diagram showing a detailed flow of the partial match search process (step S10).

  The abbreviation search unit 15 compares the search target obtained from the parse control unit 10 with the registered word of the corporate name in the dictionary storage unit 11 and searches for a range that partially matches the registered word from the back of the search target (backward Partial match search) is performed (step S101).

  The abbreviation search unit 15 proceeds to the process of step S103 when the registered word that matches the search target is not obtained (N in step S102) and obtains the registered word that partially matches the search target. (Y in step S102), the process proceeds to step S105.

  The abbreviation search unit 15 searches for a range that partially matches the registered word from the front of the search target (forward partial match search) (step S103). The abbreviation search unit 15 proceeds to the process of step S105 when a registered word that partially matches the search target (Y in step S104) and proceeds to the process of step S105 and does not obtain a registered word that matches the front part (step S104). N in S104), the result is “none” (step S106).

  The abbreviation search unit 15 identifies a registered word having “the longest matching range” or “the shortest registered word length if the matching range is the same” from the obtained result, and sets it as a result R1 (step S105). ).

  FIG. 13 is a diagram showing a detailed processing flow of the backward partial match search process (step S101), and FIG. 14 is a diagram for explaining the backward partial match search process.

  The abbreviation search unit 15 sets the reference position Dsp in the dictionary of corporate names in the dictionary storage unit 11 as the first registered word (step S1101), and extracts the registered word T at the reference position Dsp (step S1102). When the length of the registered word T (character string) is shorter than the length of the search target K (Y in step S1103), the abbreviation search unit 15 determines the character string of the registered word T as shown in the upper part of FIG. A leading position Tsp is set for the second character (step S1104). Further, the abbreviation search unit 15 extracts the character string Tstr from the beginning position Tsp to the end of the registered word (step S1105), and has the same length as the character string Tstr from the beginning of the search target K as shown in the lower part of FIG. Is extracted (step S1106).

  If the partial character string Tstr and the partial character string Kstr are the same (Y in step S1107), the abbreviation search unit 15 adds the registered word that matches the registered word T and Kstr to the result list as the matching range (step S1108). .

  If the partial character string Tstr and the partial character string Kstr are not the same (N in step S1107), the abbreviation search unit 15 checks whether the character for which the start position Tsp is set is the end of the registered word T (step S1109). . If the character for which the start position Tsp is set is the end of the registered word T (Y in step S1109), the process proceeds to step S1111. The character for which the start position Tsp is set is not the end of the registered word T (step S1109). N), the abbreviation search unit 15 shifts the head position Tsp backward by one character (step S1110), and returns to the process of step S1105.

  The abbreviation search unit 15 checks whether the reference position Dsp is the last item in the corporate name dictionary (step S1111), and if it is not the last item (N in step S1111), sets the reference position Dsp as the next item in the dictionary. If it is the last item (Y in step S1111), the result list is set as the final result (step S1113).

  FIG. 15 is a diagram illustrating a detailed processing flow of the forward partial match search in step S103.

  The abbreviation search unit 15 sets the reference position Dsp in the corporate name dictionary of the dictionary storage unit 11 as the first registered word (step S1301), and extracts the registered word T at the reference position Dsp (step S1302). If the length of the registered word T is shorter than the length of the search target K (Y in step S1303), the abbreviation search unit 15 sets the collation position Trp as the first character in the character string of the registered word T (step S1303). S1304), the collation position Krp is set to the first character of the character string of the search target K (step S1305). If the character at the collation position Trp is the same as the character at the collation position Krp (Y in step S1306), the abbreviation search unit 15 uses the character one character before the end of the registered word T as the character at the collation position Trp. It is checked whether it exists (step S1307).

  If the character at the collation position Trp and the character at the collation position Krp are not the same (N in step S1306), the abbreviation search unit 15 checks whether the character at the collation position Trp is the head of the registered word T (step S1308). ). If the character at the collation position Trp is not the head of the registered word T (N in step S1308), the process proceeds to step S1309. If the character at the collation position Trp is the head of the registered word T (Y in step S1308), step The process proceeds to S1310.

  If the character at the collation position Trp is one character before the end of the registered word T (Y in step S1307), the abbreviation search unit 15 searches the registered word that matches the registered word T, the collation position Krp from the beginning of the search target. The range up to is added as a matching range to the result list (step S1309). If the character at the collation position Trp is not the character one character before the end of the registered word T (N in step S1307), the abbreviation search unit 15 shifts the collation position Trp backward by one character (step S1310), and The collation position Krp is shifted backward by one character (step S1311), and the process returns to step S1306.

  The abbreviation search unit 15 checks whether the reference position Dsp is the last item in the corporate name dictionary (step S1312), and if it is not the last item (N in step S1312), sets the reference position Dsp as the next item in the dictionary. If it is moved (step S1313) and it is the last item (Y in step S1312), the result list is set as the final result (step S1314).

  16 and 17 are diagrams illustrating a detailed flow of the complete inclusion search process (step S11).

  The abbreviation search unit 15 sets the reference position Dsp in the corporate name dictionary of the dictionary storage unit 11 as the first registered word T (step S1501), and takes out the registered word T at the reference position Dsp ( Step S1502). If the length of the registered word T is shorter than the length of the search target K (Y in step S1503), the abbreviation search unit 15 sets the collation position Trp as the first character in the character string of the registered word T (step S1503). S1504), the collation position Krp is set to the first character of the character string of the search target K (step S1505).

  Furthermore, the abbreviation search unit 15 checks whether the character at the collation position Trp is the same as the character at the collation position Krp (step S1506). If the character at the collation position Trp and the character at the collation position Krp are the same (Y in step S1506), the process proceeds to step S1507, and if the character at the collation position Trp and the character at the collation position Krp are not the same (in step S1506). N), the process proceeds to step S1509.

  The abbreviation search unit 15 checks whether the character at the collation position Trp is the end of the registered word T (step S1507). If the character at the collation position Trp is not the end of the registered word T (N in step S1507), the abbreviation search unit 15 shifts the collation position Trp backward by one character (step S1508), and the character at the collation position Krp is: It is checked whether it is the end of the search target K (step S1509).

  If the character at the collation position Krp is not the end of the search target K (N in step S1509), the abbreviation search unit 15 shifts the collation position Krp backward by one character (step S1510), returns to the processing in step S1506, and collates the position. If the character Krp is the end of the search target K (Y in step S1509), the process proceeds to step S1513.

  If the character at the collation position Trp is the end of the registered word T (Y in step S1507), the abbreviation search unit 15 searches the registered word that matches the registered word T, the range from the head of the search target K to the collation position Krp. Is added to the result list as a matching range (step S1511).

  The abbreviation search unit 15 checks whether the reference position Dsp is the last item in the dictionary (step S1512). If the reference position Dsp is not the last item in the dictionary (N in step S1512), the abbreviation search unit 15 proceeds to the process in step S1513. If the reference position Dsp is the last item (Y in step S1512), the process proceeds to step S1514. move on. The abbreviation search unit 15 moves the reference position Dsp to the next item in the dictionary (step S1513) or selects the longest matching range from the result list as the final result (step S1514).

  As described above, the abbreviation search unit 15 of the data parsing apparatus 1 can appropriately separate the abbreviation of the corporate name and the derived name from the search target for the input corporate data, and further specify the corresponding formal name. .

  FIG. 18 and FIG. 19 are diagrams illustrating an example of parsing processing of corporate data input to the data parsing apparatus 1.

  In the example shown in FIG. 18, it is assumed that the input data is “Tomi Co., Ltd. ○ Dr. Pro Production Department”. Of the input data, the dictionary lookup unit 14 finds a registered word for the corporate case name (corporation) and the corporate name (Fujitsu) from the dictionary storage unit 11, but the department name (production unit) is not found, and the input data It shows the case where part of “Professional Production Department” becomes an unknown part.

  With the abbreviation search unit 15, for the search target “Fuji ○ Tsu Production Pro”, which is extracted from the input data by the above-described processing, “Fuku ○ Production” is the abbreviation of the corporate name, and the official name = “Fuji ○” Therefore, the parsing unit 16 identifies the “production unit” as the department name. As a result, the result output unit 17 can output the parse result shown in FIG.

  In the example shown in FIG. 19, it is assumed that the input data is “Tomi Co., Ltd. ○ Mi Miru Pro Production Department”. The dictionary lookup unit 14 finds registered words for the corporate personality name (corporation), corporate name (wealth), department name (production department) from the dictionary storage unit 11 but does not belong to them. This shows the case where “Milpro” exists in the input data and it is an unknown part.

  With the abbreviation search unit 15, for the search target “Tomi * tsumi * pro production department” extracted from the input data by the above-described processing, “Tomi * tsumi * ru” is the abbreviation of the corporate name, the official name = “ Since it is possible to search for “wealth * communicationware”, the parsing unit 16 identifies “professional production department” as the department name. As a result, the result output unit 17 can output the parse result shown in FIG.

  In FIG. 19, a part other than the legal personality name in the input data is a search target. However, a part of the search result of the department name (registered word) stored in the dictionary storage unit 11 may be used. . For example, the following procedure may be performed by using a search result having the department name as “Production Department” and setting the search range to “Tomi * tsumi * rupro”. In this case, the processing result of the abbreviation search unit 15 is “(Corporate name = corporation), Corporate name = Tomi * tsumi * ru (abbreviation), Official name = Tomi * tsumi * ware, Department name = Professional production. Part ". The parsing unit 16 can apply a combination rule of “when a plurality of department names are lined up consecutively to be combined into one department name” to make “department name = professional production department”. .

  Further, when the input data shown in FIG. 18 and FIG. 19 is processed by the conventional technique, it is considered that a result different from the processing of the disclosed data parsing apparatus 1 is obtained.

  In the input data example of FIG. 18, according to the conventional method, it is possible to discriminate the corporate name “corporation” and the corporate name “Tofu ○”, but “professional production department” and “production department” are registered words in the dictionary. Because there is no, “Professional Production Department” in the input data is an unknown part. Furthermore, even if the “Professional Production Department”, which is an unknown part, is considered an abbreviation, there is no registered word that contains all the characters of the “Professional Production Department” in the dictionary. The input data is parsed as “corporate / wealth ○ com / pro production department”.

  In the example of input data in FIG. 19, according to the conventional method, “corporation”, “wealth ○ dori”, and “production department” can be discriminated, but “Milpro” is an unknown part. Furthermore, since there is no registered word in the dictionary that contains all the characters “Milpro” for the unknown part, it cannot be divided into “Mil” and “Pro”. For this reason, “Milpro” remains as an unknown part, and it is considered that it will be divided into “Mitsubishi / Mulpro” or “Milpro / Production Department” according to the subsequent combination rules.

  Thus, it can be seen that the processing result of the data parsing apparatus 1 shown in FIGS. 18 and 19 is a more appropriate parsing result as compared with the conventional method.

  As described above, it has been explained that the data parsing device 1 can identify the abbreviations and derived names of the corporate names. However, the data parsing device 1 can search for a name other than the corporate name, for example, a department name abbreviation or Derived names can be identified.

  Further, the data parsing device 1 not only searches using the registered words stored in the dictionary storage unit 11 as they are, but also includes delimiters (for example, middle black (•), blanks, etc.) and character types included in the registered words. It is also possible to decompose the registered word using the boundary of and to use each decomposed element as a new registered word for the search.

  FIG. 20 is a diagram for explaining an abbreviation search process that uses a registered word in the department name dictionary after being decomposed.

  It is assumed that a department name dictionary is stored in the dictionary storage unit 11 and a department name “Date Management and Firmware Division” is registered in the department name dictionary. Further, it is assumed that there is a search target “data mill process meeting” extracted from input data.

  The abbreviation search unit 15 recognizes the boundary between the character “Nakuguro” (·) appearing in the registered word and the character type, and the registered words are “registered data”, “Mileware”, and “business department”. Into divided registration words. That is, in the case of the department name dictionary shown in FIG. , Management department "is treated as having a divided registered word.

  Similar to the corporate name abbreviation search process, the abbreviation search unit 15 assigns the divided registration word “data XX management” of “data” to the search target “data mile event process meeting”. Search the divided registration word "Milware", search the divided registration word "Division" for "thing" as the most similar registered word, and search for "data", "mix", and "thing". It is determined as an abbreviation of the searched registered word. The parsing unit 16 can divide the search target into “data”, “mix”, “thing”, and “schedule meeting”, and the result output unit 17 can search the search target “data mix process meeting”. For example, it is possible to output a perspective result of a first department name “Day XX Management and Wareware Division” and a second department name “Process Meeting”.

  In the conventional method, since the department name that exactly matches the department name dictionary is not registered, the search target “data mill event meeting” is output without being divided. However, it can be seen that the data parsing apparatus 1 can output a parsing result that seems to be more appropriate for corporate data including the “data mill process meeting”.

  FIG. 21 is a diagram illustrating a hardware configuration example for implementing the data parsing device 1.

  The data parsing device 1 includes a computer 100 having an arithmetic unit (CPU) 101, a temporary storage device (DRAM, flash memory, etc.) 102, a permanent storage device (HDD, flash memory, etc.) 103, and an input device (keyboard, mouse, etc.). ) 120 and an output device (display, printer, etc.) 130.

  The data parsing device 1 can be implemented by a program that can be executed by the computer 100. In this case, a program describing the processing contents of the functions that the data parsing apparatus 1 should have is provided, and the computer 100 executes the provided program, whereby the processing unit of the data parsing apparatus 1 described above is executed by the computer 100. Realized above.

  That is, each processing unit is realized by a program for functioning as each processing unit included in the data parsing device 1 being loaded into the temporary storage device 102 and executed by the arithmetic unit (CPU) 101. Various information used by the data parsing device 1 is stored in the persistent storage device 103.

  The computer 100 directly reads the above program from a portable recording medium and executes the processing according to the program or receives the transferred program sequentially each time the program is transferred from another computer. Can also be executed.

  Further, the program can be recorded on a recording medium readable by the computer 100.

  As described above, according to the abbreviation search unit 15 of the disclosed data parsing device 1, even if the company data includes a name such as a company name or a department name in the form of an abbreviation or a derived name, the official name A part that is regarded as an abbreviated name and a derived name can be identified using a dictionary in which only the registered abbreviations are registered, and a formal name corresponding to the identified abbreviated name and derived name can be searched.

  Further, according to the disclosed data parsing device 1, the processing result of the abbreviation search unit 15 is used to divide the corporation data inputted by treating the character string identified as the abbreviation and the part regarded as the derived name as one element. In addition, a parsing result in which the identified abbreviation and the formal name corresponding to the derived name are associated can be output.

  Therefore, in the parsing processing of corporate data that is a character string including abbreviations and derived names, the corresponding or related official names can be specified, and the accuracy of subsequent name identification processing and data cleansing processing can be improved. , It can contribute to reducing the burden of checking work.

  In the above description, the case where the present invention is applied to the parsing process of corporate data by referring to the data including the abbreviation and the derived name of the corporate name, which is the technical field as the background, as corporate data. In the above description, corporate data means text data including an expression indicating an abbreviated name or a derived name of the official name, and the present invention is not limited to the above description, Naturally, various modifications are possible in the range.

  The features in the embodiments of the present invention are listed as follows.

(Appendix 1)
A dictionary storage unit for storing a name dictionary whose official name is a registered word;
A data input part that accepts input of a character string to be searched;
When the search target character string includes a portion that does not match a registered word in the name dictionary, a search is performed for a registered word that partially matches the search target from the name dictionary, and among the searched registered words, A registered word that has a maximum range that partially matches the search target and that has a maximum percentage of the range that matches the registered word is determined, and the search target portion is determined as an abbreviation, and the specified And a search processing unit that uses the registered word as a formal name of the abbreviation.

(Appendix 2)
When the search target character string includes a portion that does not match a registered word in the name dictionary, the search processing unit is completely included in the search target from the name dictionary and has the maximum length. 2. The abbreviation search device according to appendix 1, wherein a registered word is specified, the portion to be searched is determined as a derived name, and the specified registered word is an official name related to the derived name.

(Appendix 3)
The abbreviation search device according to Supplementary Note 1 or Supplementary Note 2, wherein the official name stored in the dictionary storage unit is a legal name of a corporation.

(Appendix 4)
In the case where a corporate case name dictionary having a corporate case name as a registered word is stored in the dictionary storage unit, the search processing unit removes the corporate case name registered in the corporate case name dictionary from the input character string. The abbreviation search device according to Supplementary Note 3, wherein a character string is the search target.

(Appendix 5)
The search processing unit, when a department name dictionary having a department name as a registered word is stored in the dictionary storage part, excludes the department name registered in the department name dictionary from the input character string. The abbreviation search device according to Supplementary Note 3 or Supplementary Note 4, wherein a character string is the search target.

(Appendix 6)
A processing method executed by a computer having a dictionary storage unit,
A processing step for accepting input of a character string to be searched;
Referring to the dictionary storage unit that stores a name dictionary with a formal name as a registered word, when the character string to be searched includes a portion that does not match a registered word in the name dictionary, the name to be searched from the name dictionary Processing steps to search for registered words that partially match
Among the registered words that have been searched, a registered word that has the largest range that partially matches the search target and that has the largest percentage of the range that matches the registered word is identified, and the search target portion And a processing step of determining the registered word as the formal name of the abbreviation.

(Appendix 7)
In a computer with a dictionary storage,
A process for accepting input of a character string to be searched;
Referring to the dictionary storage unit that stores a name dictionary with a formal name as a registered word, when the character string to be searched includes a portion that does not match a registered word in the name dictionary, the name to be searched from the name dictionary Searching for registered words that partially match
Among the registered words that have been searched, a registered word that has the largest range that partially matches the search target and that has the largest percentage of the range that matches the registered word is identified, and the search target portion An abbreviation search program for executing a process for determining a registered abbreviation as a formal name of the abbreviation.

(Appendix 8)
A dictionary storage unit for storing a name dictionary whose official name is a registered word;
A data input part that accepts input of a character string to be searched;
When the search target character string includes a portion that does not match a registered word in the name dictionary, a search is performed for a registered word that partially matches the search target from the name dictionary, and among the searched registered words, A registered word that has a maximum range that partially matches the search target and that has a maximum percentage of the range that matches the registered word is determined, and the search target portion is determined as an abbreviation, and the specified A search processing unit that uses the registered word as the formal name of the abbreviation,
A parsing processing unit that divides the search target character string into respective names that are arranged according to the combination rule based on a combination rule indicating the arrangement for each name;
The data parsing device having an abbreviation search function, wherein the parsing processing unit divides a portion determined as the abbreviation as one name and associates the specified formal name with the abbreviation.

(Appendix 9)
When the search target character string includes a portion that does not match a registered word in the name dictionary, the search processing unit is completely included in the search target from the name dictionary and has the maximum length. Identifying a registered word, determining that the part to be searched is a derived name, and performing the process of making the identified registered word an official name related to the derived name;
9. The abbreviation search according to claim 8, wherein the parse processing unit divides the part determined as the derived name as one name, and performs processing for associating the specified formal name with the derived name. Data parsing device with functions.

(Appendix 10)
The data parsing device having the abbreviation search function according to the supplementary note 8 or the supplementary note 9, wherein the search processing unit has a formal name stored in the dictionary storage unit as a legal name of a corporation.

(Appendix 11)
The search processing unit, when a corporate case name dictionary having a corporate case name as a registered word is stored in the dictionary storage unit, characters obtained by removing the corporate case name registered in the corporate case name dictionary from the input character string The data parsing apparatus having the abbreviation search function according to Supplementary Note 10, wherein a column is the search target.

(Appendix 12)
The search processing unit, when a department name dictionary having a department name as a registered word is stored in the dictionary storage part, excludes the department name registered in the department name dictionary from the input character string. The data parsing device having the abbreviation search function according to the supplementary note 10 or the supplementary note 11, wherein a character string is the search target.

DESCRIPTION OF SYMBOLS 1 Data parsing apparatus 10 Parsing control part 11 Dictionary memory | storage part 12 Combination rule memory | storage part 13 Data input part 14 Dictionary lookup part 15 Abbreviation search part 16 Parsing part 17 Result output part

Claims (9)

A storage unit for storing a plurality of registered official names ;
A data input part that accepts an input string;
A processing unit,
The processor is
When the input string does not match any of the plurality of registered official name, from said storage unit, formally a string containing a character string that matches the input character string and partly are the plurality of registered Search by name ,
In the search results extracted official name, a maximum length of the string that matches the input character string and partly, and the string to the partial match to the length of the official name Identify the official name with the largest percentage of length,
The partially matching character string is determined as a character string indicating an abbreviation corresponding to the specified official name .
A search device characterized by that. The search device according to claim 1, wherein the plurality of registered formal names stored in the storage unit include at least one of a corporate name, a corporate case name, and a department name. When the input character string does not completely match the official name stored in the storage unit, the processing unit is completely included in the input character string from the storage unit, and the length of the character string is maximum. identify the official name string is, the input string, and characterized in that determines a derived name corresponding to the specified the official name, and full name associated a formal name that the particular to the derivation name The search device according to claim 1 or 2. The search device according to any one of claims 1 to 3, wherein the processing unit has a formal name stored in the storage unit as a legal name of a corporation. A processing method executed by a computer having a storage unit,
Processing steps for accepting input strings;
Referring to the storage unit that stores a plurality of registered formal names , if the input character string does not match any of the plurality of registered formal names , the input character string and a part are stored from the storage unit. A processing step of searching the plurality of registered formal names for a character string that includes a character string that matches automatically,
In the search results extracted official name, a maximum length of the string that matches the input character string and partly, and the string to the partial match to the length of the official name And a processing step of identifying a formal name having a maximum length ratio and determining the partially matching character string as a character string indicating an abbreviation corresponding to the identified formal name. retrieval method. To a computer with a storage unit,
Processing to accept an input string;
Referring to the storage unit that stores a plurality of registered formal names , if the input character string does not match any of the plurality of registered formal names , the input character string and a part are stored from the storage unit. A process of searching for a character string including a character string that matches automatically from the plurality of registered official names ;
In the search results extracted official name, a maximum length of the string that matches the input character string and partly, and the string to the partial match to the length of the official name A search program for identifying a formal name having a maximum length ratio and determining the partially matched character string as a character string indicating an abbreviation corresponding to the identified formal name . A parsing processing unit for executing a parsing process for dividing a character string having a plurality of names into constituent elements according to a combination rule indicating a correct arrangement order of the plurality of names as a character string structure. A data parsing device comprising:
A storage unit for storing a name dictionary whose official name is a registered word;
A data input part that accepts an input string;
When the input character string includes a portion that does not match a registered word in the name dictionary, the registered word in the name dictionary is compared with the input character string, and the registered word and a portion that partially matches the registered word A character string is extracted in association with one another, and from the extracted registered words, one registered word having the maximum length of the associated partial character string and the maximum ratio to the input character string is identified, and the input A partial character string associated with the specified registered word of the character string is determined as an abbreviation, and a search processing unit that uses the specified registered word as the formal name of the abbreviation,
The parsing processing unit divides a part of the input character string determined as the abbreviation in the parsing process as one name, and performs a process of associating the specified formal name with the abbreviation. Data parsing device with functions. The search processing unit, when the input character string includes a portion that does not match a registered word of the name dictionary, is registered from the name dictionary completely in the input character string and has the maximum length. Identifying a word, determining that the portion of the input character string is a derived name, and performing a process of making the identified registered word a formal name related to the derived name;
The search function according to claim 7, wherein the parsing processing unit divides the portion determined as the derived name as one name, and performs processing for associating the specified formal name with the derived name. A data parsing device. The data parsing apparatus having a search function according to claim 7 or 8, wherein the search processing unit has a formal name stored in the storage unit as a legal name of a corporation.

Images