JP4139805B2 - Apparatus, method and program for converting lexical data to data - Google Patents

Description

  The present invention relates to an apparatus, a method, and a program for converting a lexical phrase including a plurality of words into another data including one piece of information specified by the plurality of words.

  For example, a bank or the like may require a customer information (CIF) analysis process using a computer for a settlement instruction message received (CIF means a customer information file). .) Specifically, for example, a combination of a plurality of words “GETRONICS”, “FOODS”, “CO., LTD”, and “AKASAKA” in the word “GETRONICS FOODS CO., LTD 1-2-34 AKASAKA” It is necessary to convert the customer code (that is, the CIF code) to “1234567”.

  Conventionally, this conversion processing is performed as follows. That is, a customer code table of a combination of a plurality of words specifying a customer code and a corresponding customer code is stored in the storage device in advance. Next, a plurality of words to be converted are extracted from the input data, and the extracted plurality of words and a plurality of words in the customer code table are subjected to character string comparison. (See Patent Document 1 for such an example.)

  However, since such character string comparison is performed for each character (in 1-byte units), especially when a large amount of data is targeted, the processing is performed at high speed due to the search time required for the processing. There was a problem that could not be done.

  As a method for overcoming such a problem, it is possible to symbolize a dictionary used for converting a lexical in a message into a CIF code on a main memory using a memory address as a symbol. It was proposed in a patent application (see Patent Document 2). Specifically, a basic word dictionary associated with each of a plurality of words and their attribute information and registered in the secondary storage device in advance, a combination of two words registered in the basic word dictionary, and a key The nuclear name dictionary that is pre-registered in the secondary storage device in association with the selected word in the set and the symbol specified by the set, the symbol and another word not included in the set, and the The full name dictionary registered in advance in the secondary storage device in association with the symbol and the CIF code specified by another word is symbolized on the main memory using the memory address as a symbol. By referring to the symbolized basic word dictionary, an address corresponding to the word included in the converted lexical phrase is acquired. A pair that matches two addresses (one of which is a key) is found from the symbolized nuclear name dictionary, and then a pair that matches the other address is found from the symbolized full name dictionary, The conversion token is converted into a target CIF code.

  As described above, this method is a specification in which a key (symbol) that associates each dictionary is directly assigned to a physical memory address. When such a specification is used, the memory address is dynamically allocated as a characteristic of the computer. Therefore, each time the program is loaded into the memory, the key relationship between the dictionaries is changed to the acquired memory address. It is necessary to replace and generate, and in this respect, high load and required time are required in the program initial processing.

  As a method for overcoming such a problem, a logical index representing each word is allocated in advance, and the word and the corresponding logical index are stored in a dictionary in association with each other, and this is expanded in a memory. It was proposed in patent applications prior to this patent application (see Patent Documents 3 and 4).

  In addition, when performing an analysis process for converting a word into a logical index, some words to be analyzed have the same semantic content in the analysis process, although the notation is different. Although different logical indexes may be assigned to these words individually, they have the same semantic content in the analysis process, so even if the same logical index is assigned, no problem occurs in the analysis process, and there is a waste in processing efficiency. It can be omitted. Therefore, it has been proposed in patent applications prior to this patent application to store a plurality of words having the same semantic content in the analysis process although they are different in notation in one line of the word column of the basic word dictionary. (See Patent Document 4).

  Furthermore, when converting a lexical into data, in order to improve the accuracy of analysis processing so that one lexical can be correctly converted even if there is an inclusion relationship that includes the other lexical, A set of logical index sets and their corresponding CIF codes are stored in the name dictionary and are not included in the included logical index set, so they are not used directly to identify CIF codes, but identify CIF codes A logical index corresponding to an unnecessary word, which is a word that assists the user, is stored in the unnecessary word dictionary in association with the CIF code, and is converted into a set of lexical logical indexes including unnecessary words. The CIF code is specified by comparing (partial matching) with the logical index set of the name dictionary including the logical index. Proposed in prior patent application from the patent application (see Patent Document 5).

In addition, a method for shortening access time by attaching an index to a word on a dictionary and associating it with a storage position on a disk is known (see, for example, Patent Document 6).
JP 2002-56005 A Patent application for Japanese Patent Application No. 2003-115287 Patent application for Japanese Patent Application No. 2004-89101 Patent application No. 2004-89121 Patent application for Japanese Patent Application No. 2004-274422 JP-A-7-36900

  As described above, when storing a plurality of words having different meanings but having the same meaning in the analysis process in one line of the word column of the basic word dictionary (see Patent Document 4), specifically, Requires an operator to input a plurality of words and corresponding logical indexes from the input means. Inputting a plurality of words and corresponding logical indexes is likely to cause errors because of the large amount of input. The problem is that even if a word and a corresponding logical index are already registered in the basic word dictionary and another word having the same meaning in the analysis processing is added although the notation is different, The corresponding logical index must be entered. Therefore, even when the information is added, the operator's input amount is still large and errors are likely to occur.

  Therefore, an object of the present invention is to reduce the amount of data that an operator must input when adding words having different meanings but having the same meaning in the analysis process to the basic word dictionary as much as possible to reduce errors. It is to generate a highly reliable basic word dictionary database.

The above-described problem is an apparatus for converting a lexical phrase including a plurality of words into another data including one piece of information specified by the plurality of words according to one aspect of the present invention, and is stored in advance in a storage unit. A word field for storing words and a logical index field for storing a logical index representing the words stored in the word field, wherein each of the plurality of words is associated with a logical index representing each word. The plurality of word fields are registered in advance, and the word column has the same logical index attached to a plurality of words having the same semantic content and different notation in the analysis process for converting the word into a logical index. Corresponds to a basic word dictionary configured to be able to store words and a set of words that are pre-stored in the storage means and combined with words registered in the basic word dictionary The logical index set is the same as the name dictionary registered in advance in association with one piece of information specified by the word set, and the semantic content in the analysis process registered in the basic word dictionary is the same. A plurality of words having different notations are associated with the same logical index assigned to the plurality of words having different notations and expanded in memory, and a set of logical indexes registered in the name dictionary The memory is expanded in the memory in association with one piece of information specified by the set of words corresponding to each set of logical indexes, the lexical phrase to be converted including a plurality of words is decomposed into words, Referring to the basic word dictionary expanded in memory, a logical index corresponding to the decomposed word is obtained, and the memory-expanded logical index set is obtained A logical index set that matches any logical index set composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed word is selected, and 1 specified by the selected logical index set. By using one piece of information, the lexical phrase including the plurality of words is converted into another data including the one specified information, and the logical index corresponding to the decomposed word is obtained. If there is a word that cannot be acquired, the word is specified, and the specified word is stored in the analysis NG logical index set file stored in the storage unit together with the logical index of the word that has acquired the logical index among the decomposed words. First processing engine to store and input for entering information including data And at least the corresponding logical index could not be acquired from the analysis NG logical index set file in response to a search command input from the input means and display means for displaying information including data A word and a logical index stored together with the word are searched and displayed on the display means, and the logical index displayed on the display means inputted from the input means is not stored and stored together with the word. The name dictionary is searched using one piece of information estimated from the logical index as a key, a set of logical indexes corresponding to the one piece of estimated information is obtained, the word for which the logical index could not be obtained, and the A logical index stored with the word and a logical index obtained from the name dictionary. A second processing engine that identifies a difference word and a logical index with a set of pairs and registers the difference word as a synonym in the word field in the basic word dictionary corresponding to the difference logical index; It is solved by the apparatus characterized by comprising.

The above object is also achieved by a separate in accordance with an aspect, an apparatus for converting the other data including one information specified lexical including a plurality of words by the plurality of words of the present invention, advance in the storage means A word field for storing words and a logical index field for storing a logical index representing the words stored in the word field, and each of the plurality of words is a logical index representing each word; The word column is registered in advance, and the word column is assigned the same logical index to a plurality of words having the same meaning and different notation in the analysis process for converting the word into a logical index. A basic word dictionary configured to be capable of storing a plurality of words, and a set of words obtained by combining words that are stored in the storage unit in advance and registered in the basic word dictionary A set of corresponding logical indexes displays a name dictionary registered in advance in association with one piece of information specified by the set of words, an input means for inputting information including data, and information including data A plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis process and are assigned to the different words with the notation And a logical index set registered in the name dictionary is associated with one piece of information specified by the word set corresponding to each logical index set. Expand the memory on the memory, disassemble the lexical word to be converted including a plurality of words into words, refer to the basic word dictionary expanded in the memory on the memory, A logical index corresponding to the solved word is obtained, and any one of the logical indexes composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed word is selected from the set of logical indexes expanded in the memory. By selecting a set of logical indexes that matches the set and using one piece of information specified by the set of selected logical indexes, another phrase containing the one piece of information that is specified When converting to data and obtaining a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be obtained, the word is identified, and the identified word is the logical index of the decomposed word Is displayed on the display means together with the logical index of the word that has been acquired, and the input means The name dictionary using, as a key, one piece of information estimated from the logical index of the word for which the logical index displayed in the display means and the word for which the logical index has been acquired among the decomposed words. To obtain a set of logical indexes corresponding to the one piece of estimated information, and a logical index of a word for which a logical index could be obtained among the words for which the logical index could not be obtained and the decomposed words, , Specifying a difference word and a logical index from a set of logical indexes acquired from the name dictionary, and registering the difference word as a synonym in the word field in the basic word dictionary corresponding to the difference logical index And a processing engine .

Furthermore, the subject is a method of converting a lexical phrase including a plurality of words into another data including one piece of information specified by the plurality of words, according to one aspect of the present invention, and storing the words It has at least a logical index field for storing a word field and a logical index representing the word stored in the word field, and each of the plurality of words is registered in advance in association with a logical index representing each word. In addition, the word column can store the plurality of words so that the same logical index is attached to a plurality of words having the same semantic content and different notation in the analysis processing for converting the word into a logical index. A step of storing the constructed basic word dictionary in the storage means, and a logical input corresponding to a set of words obtained by combining words registered in the basic word dictionary A step of storing a name dictionary registered in advance in association with one piece of information specified by the set of words in the storage means, and a step in the analysis process registered in the basic word dictionary Associating a plurality of words having the same semantic content but different notations with the same logical index assigned to the plurality of different notations in the memory, and registering them in the name dictionary Associating a set of logical indexes associated with one piece of information specified by the set of words corresponding to each set of logical indexes into the memory, and a lexical phrase to be converted comprising a plurality of words Decomposing into words, and referring to a basic word dictionary expanded in the memory, a logical index corresponding to the decomposed words is obtained. And a logical index set that matches one of the logical index sets of a plurality of logical indexes among the logical indexes corresponding to the decomposed word from the set of logical indexes expanded in the memory. Converting the lexical phrase including the plurality of words into another data including the specified one information by using one information specified by the selected set of logical indexes, and When acquiring a logical index corresponding to a decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word is the logic of the word that has acquired a logical index among the decomposed words. The analysis NG logical index set file stored in the storage means together with the index And a word for which at least the corresponding logical index could not be obtained from the analysis NG logical index set file in response to a search command input from an input means for inputting information including data A logical index stored together with the word, and displaying on the display means for displaying information including data; and obtaining the logical index displayed on the display means, which is input from the input means Searching the name dictionary using as a key one word estimated from a word that could not be performed and a logical index stored together with the word, and obtaining a set of logical indexes corresponding to the one estimated information; , The word for which the logical index could not be obtained and stored together with the word A difference word and a logical index between a logical index and a set of logical indexes acquired from the name dictionary are identified, and the difference word is synonymous with the word field in the basic word dictionary corresponding to the difference logical index And a step of registering as a word .

Furthermore, the above object is achieved by a separate in accordance with an aspect, a method of converting a lexical including a plurality of words to another data including one of the information specified by the plurality of words of the present invention, a word At least a logical index field for storing a word field to be stored and a logical index representing the word stored in the word field, and each of the plurality of words is pre-registered in association with a logical index representing each word In addition, the word column stores the plurality of words so that the same logical index is attached to a plurality of words having the same semantic content and different notation in the analysis processing for converting the word into a logical index. A basic word dictionary configured to be stored in a storage means, and a theory corresponding to a set of words obtained by combining words registered in the basic word dictionary A step of storing a name dictionary registered in advance in association with one piece of information specified by the set of words as a set of indexes, and a step of storing in the basic word dictionary Associating a plurality of words having the same semantic content but different notations with the same logical index assigned to the plurality of different notations in the memory, and registering them in the name dictionary Associating a set of logical indexes associated with one piece of information specified by the set of words corresponding to each set of logical indexes into the memory, and a lexical phrase to be converted comprising a plurality of words A logical index corresponding to the decomposed word with reference to a basic word dictionary expanded in the memory; And a logic that matches one of the logical index sets composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed words from the set of logical indexes expanded in the memory. Selecting an index set and using one piece of information specified by the selected set of logical indexes to convert a lexical phrase including the plurality of words into another data including the specified one piece of information. When acquiring a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word can be acquired from the decomposed word. A word displayed on the display means for displaying information including data together with a logical index of words. And the logic of the word that can be obtained from the input means for inputting information including data, the word that could not obtain the logical index displayed on the display means, and the word that could obtain the logical index among the decomposed words Searching the name dictionary using one piece of information estimated from the index as a key, obtaining a set of logical indexes corresponding to the one piece of estimated information, the word that could not obtain the logical index, and the The basic word corresponding to the logical index of the difference is specified by identifying a difference word and a logical index between the logical index of the word that has obtained the logical index among the decomposed words and the set of logical indexes acquired from the name dictionary Registering the difference word as a synonym in the word field in the dictionary; It is solved by a method characterized by comprising.

According to still another aspect of the present invention, the above-described problem has at least a word column for storing words and a logical index column for storing logical indexes representing words stored in the word columns, Each is registered in advance in association with a logical index representing each word, and the word field has a plurality of different meanings and the same semantic content in the analysis process for converting the word into a logical index. A basic word dictionary configured to be able to store the plurality of words so that the same logical index is attached to the word, and a logical index corresponding to a set of words obtained by combining words registered in the basic word dictionary A storage unit that stores a name dictionary registered in advance in association with one piece of information specified by the set of words, and information including data is input. A computer comprising an input means for displaying and a display means for displaying information including data, wherein a lexical phrase including a plurality of words is converted into another data including one piece of information specified by the plurality of words. A program for causing a computer to function as a conversion device, wherein a plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis process are different in the notation. A set of logical indexes associated with the same logical index assigned to a plurality of words is expanded in a memory and registered in the name dictionary. The memory is expanded on the memory in association with one piece of information specified by the above, the lexical phrase to be converted including a plurality of words is decomposed into words, and the memo The logical index corresponding to the decomposed word is obtained by referring to the basic word dictionary expanded in the memory, and the logical index corresponding to the decomposed word is selected from the set of logical indexes expanded in the memory. By selecting a logical index set that matches any logical index set composed of a plurality of logical indexes, and using one piece of information specified by the selected logical index set, the plurality of words are Converting the lexical word to be included into another data including the one specified information, and when obtaining a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be obtained, identify the word; For the identified word, the logical index of the word for which the logical index of the decomposed word can be obtained. And a first processing engine stored in the analysis NG logical index set file stored in the storage means together with the search, and in response to the search command input from the input means, from the analysis NG logical index set file At least the corresponding logical index could not be obtained and the logical index stored together with the word are searched and displayed on the display means, and the logical index displayed on the display means input from the input means Search the name dictionary using as a key one information estimated from a word that could not be acquired and a logical index stored together with the word, and obtain a set of logical indexes corresponding to the one estimated information, The logical index could not be obtained and stored together with the word A difference word and a logical index between a logical index and a set of logical indexes acquired from the name dictionary are specified, and the difference word is synonymous with the word field in the basic word dictionary corresponding to the difference logical index. This is solved by a program characterized by functioning as a second processing engine registered as a word .

According to another aspect of the present invention, the above-described problem includes at least a word column for storing words and a logical index column for storing logical indexes representing words stored in the word columns, and each of a plurality of words Are registered in advance in association with a logical index representing each word, and the word field has a plurality of words having the same semantic content in the analysis processing for converting the word into a logical index and different notations A basic word dictionary configured to be able to store the plurality of words so that the same logical index is attached to the same, and a set of logical indexes corresponding to a set of words obtained by combining words registered in the basic word dictionary Inputs a storage means for storing a name dictionary registered in advance in association with one piece of information specified by the set of words, and information including data A computer comprising input means for displaying and display means for displaying information including data, and converting a lexical phrase including a plurality of words into another data including one piece of information specified by the plurality of words A program for causing a computer to function as a plurality of words having the same meaning and different notation in the analysis process registered in the basic word dictionary. In association with the same logical index assigned to the word, the memory is expanded in the memory, and the set of logical indexes registered in the name dictionary is determined by the set of words corresponding to each set of logical indexes. The memory is expanded on the memory in association with one specified information, the lexical phrase to be converted including a plurality of words is decomposed into words, A logical index corresponding to the decomposed word is obtained by referring to a basic word dictionary expanded in memory, and a plurality of logical indexes corresponding to the decomposed word are selected from the set of logical indexes expanded in memory. A lexical phrase including the plurality of words is selected by selecting a logical index set that matches any logical index set composed of logical indexes and using one piece of information specified by the selected logical index set. Is converted into another data including the specified one information, and when obtaining a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be obtained, the word is identified, and the identification is performed. The logical index of the word that was able to obtain the logical index among the decomposed words In addition to the words that are displayed on the display means and input from the input means, the logical index displayed on the display means that could not be acquired, and the logical index of the words that were able to acquire the logical index among the decomposed words The name dictionary is searched using one piece of information as a key to obtain a set of logical indexes corresponding to the one piece of estimated information, and among the words that could not obtain the logical index and the decomposed words A word and a logical index of a difference between a logical index of a word for which a logical index can be obtained and a set of logical indexes obtained from the name dictionary are specified, and the word in the basic word dictionary corresponding to the logical index of the difference Function as a processing engine for registering the difference word as a synonym in the column. It is solved by a program and symptoms.

  According to one aspect of the present invention, the number of difference words is preferably one.

  In the present invention, with the configuration as described above, an operator simply inputs one piece of information specified by a plurality of words for a lexical phrase including a plurality of words, and the logical index corresponding to the word in the word column corresponds to the logical index column. Is automatically registered in the word field where the word is registered in the basic word dictionary in which the word is already registered and has the same meaning content in the analysis processing as the word, so the operator should input It is possible to generate a highly reliable basic word dictionary database with a data amount as small as possible and therefore with few errors.

Preferred embodiments of the present invention will be described below with reference to the drawings. Throughout the drawings, the same or similar reference numerals indicate the same or similar components.
First, with reference to Patent Document 3, an apparatus for converting a lexical phrase using the present invention into data will be described. FIG. 10 is a diagram illustrating a basic configuration of an example of an apparatus that converts a lexical phrase used for performing customer customer information (CIF) analysis processing using a computer to a payment instruction message received at a bank or the like into data, for example. It is. In FIG. 10, 10 is a processing engine composed of an arbitrary data processing device such as a main frame computer, personal computer, microprocessor or the like, 12 is a main memory, 14 is a basic word dictionary, and 16 is a name. Each dictionary is shown. The basic word dictionary 14 and the name dictionary 16 are stored in a hard disk (not shown) such as a magnetic disk, but are not limited thereto, and can be stored in any other type of storage device. The data processing device functioning as the processing engine 10 and the main memory 12, the hard disk storing the basic word dictionary 14 and the name dictionary 16 are coupled to each other by a normal data bus or the like (not shown). Yes.

  FIG. 11 shows a state in which words registered in advance in the basic word dictionary 14 (hereinafter also referred to as “basic words”) are expanded in the main memory 12. The input data handled in the CIF analysis processing in a bank or the like is usually composed of alphabets and numerics (numeric characters), and therefore the words registered in the basic word dictionary 14 are also composed of alphabets and numerics (numeric characters). Note that the word or basic word in this specification includes not only common nouns, proper nouns, and abbreviations, but also any set of symbols having a certain meaning. As shown in FIG. 11, an example of the basic word dictionary 14 includes words, parts of speech, name attributes, code attributes, and logical indexes represented by numbers corresponding to the registered words. The number of the logical index is preferably a number expressed in binary to reduce the number of CPU processing units. The basic word dictionary should include at least a word column for registering a basic word and a logical index column for storing a logical index of a number corresponding to the word. Other items are limited to the above. Is not to be done. When a basic word is registered in the word column of the basic word dictionary 14, the logical index is assigned a number (fixed value) corresponding to the word, and is registered in the logical index column of the basic word dictionary 14. The number is also stored, that is, registered in association with. When the basic word dictionary 14 includes a basic word that is not registered in the lexical phrase to be converted, the basic word dictionary 14 registers a new basic word and a corresponding logical index, and when the basic word dictionary 14 is no longer used with the registered basic word, the basic word dictionary 14 The structure is preferably such that a word and its corresponding logical index can be deleted.

  The processing engine 10 expands the basic words registered in the basic word dictionary 14 in the main memory 12 as shown in FIG. 11 before starting the conversion process. Since memory addresses are dynamically allocated, the same word is not stored at the same memory address location each time a program is memory loaded, and the memory addresses are typically different. Thus, in this example, for the basic word “AKASAKA” in the word column, as indicated by reference numeral 100, on the main memory 12 that is dynamically allocated when the program is memory loaded at a given time. Number 50 is assigned as the memory address, and information relating to “AKASAKA” is stored at that location. Note that the basic word “AKASAKA” is stored in the main memory 12 in association with a fixed logical index “# 100”, in contrast to the memory address. From the viewpoint of high-speed processing, it is preferable that all the basic word dictionaries 14 are expanded in the main memory 12, but this is used when the basic words included in the lexical terms to be converted are known in advance. Only basic words may be expanded in the main memory 12, and depending on the application, the conversion processing speed may be slow, but a part of the basic words registered in the basic word dictionary 14 is expanded in the memory. When the basic word of expansion becomes necessary for the conversion process, the memory may be additionally expanded.

  FIG. 12 shows a state in which a logical index set registered in advance in the name dictionary 16 is expanded on the main memory 12. FIG. 12 also shows a state in which basic words registered in advance in the basic word dictionary 14 similar to FIG. 11 are expanded in the main memory 12. As shown in FIG. 12, the item of the name dictionary 16 includes one piece of information specified by a combination of a logical index set, a name, and a word indicated by the name, for example, a CIF code in this bank application example. Among the basic words registered in the basic word dictionary 14, basic words included in combinations of basic words for which the CIF code is specified are registered in advance in the name column of the name dictionary 16. As an example, the CIF code “123-45678” is obtained by the word combination “GETRONICS FOODS CO., LTD AKASAKA”, “101-23456” is obtained by “GETRONICS SHOKUHIN OSAKA”, and “102-34” is obtained by “GETRONICS BANK AKASAKA 56”. Assume that each is specified. Therefore, “GETRONICS FOODS CO., LTD AKASAKA” is registered in the first line of the name dictionary 16, “GETRONICS SHOKUHIN OSAKA” is registered in the second line, and “GETRONICS BANK AKASAKA” is registered in the third line. Has been. In the column of logical index set, a set of logical indexes corresponding to these combined basic words is registered. For example, “# 107, # 106, # 104, # 100” corresponding to the name “GETRONICS FOODS CO., LTD AKASAKA” is displayed in the first line of the name dictionary 16, and the name “ “# 107, # 112, # 111” corresponds to “GETRONICS SHOKUHIN OSAKA”, and “# 107, # 101, # 100” corresponds to the name “GETRONICS BANK AKASAKA” in the third row, respectively. It is registered as a set of logical indexes of the basic word dictionary 14. A unique number corresponding to each name is registered in the CIF code column. For example, the first line of the name dictionary 16 has a CIF code “123-45678” corresponding to “GETRONICS FOODS CO., LTD AKASAKA”, and the second line has a CIF code “123-5678” corresponding to “GETRONICS SHOKUHIN OSAKA”. “101-23456” is registered in the third row, and the CIF code “102-34567” corresponding to “GETRONICS BANK AKASAKA” is registered. The name dictionary 16 can be deleted when there is a combination of basic words that are not registered in the lexical phrase to be converted, and a combination that includes a new basic word is registered and can be deleted when the registered combination is no longer used. It is preferable that

  Before starting the conversion process, the processing engine 10 associates the logical index set registered in the name dictionary 16 with the CIF code, and stores the memory on the main memory 12 as indicated by reference numeral 16a in FIG. expand.

  Next, input data conversion processing will be described with reference to FIGS. FIG. 13 is a diagram for explaining a process in which data input to the conversion device shown in FIG. 10 is converted.

  Here, it is assumed that the basic word dictionary 14 and the name dictionary 16 are expanded on the main memory 12 as described above. Then, it is assumed that data indicated by reference numeral 40 in FIG. 13 is input. The processing engine 10 breaks the input data 40 into words as shown in step 42. In this example, as indicated by reference numeral 40, the input data is composed of alphabets and numerics (numeric characters), and the words are separated by spaces. The processing engine 10 breaks down the input data 40 into words with a space included in the input data 40 character string as a delimiter.

  Next, as shown in step 44, the processing engine 10 acquires the logical index corresponding to the decomposed word with reference to the basic word dictionary 14a expanded in the memory on the main memory 12 shown in FIG. . In this example, the logical index corresponding to the basic word circled in the basic word dictionary 14a of FIG. 12 is acquired.

  Next, in step 46, the processing engine 10 converts the word that has been obtained with the logical index out of the decomposed words into the logical index that has been obtained. If <1-234> does not exist in the basic word dictionary 14a, it is left as it is. In this example, after conversion, logical indexes “# 107”, “# 106”, “# 104”, and “# 100” are obtained.

  In step 48, the processing engine 10 stores an arbitrary set of logical indexes “# 107”, “# 106”, “# 104”, and “# 100” in the name dictionary 16a expanded in the memory shown in FIG. A search is made as to whether there is a logical index set. If they match, a CIF code corresponding to the logical index set is obtained. In this example, the combinations of logical indexes “# 107”, “# 106”, “# 104”, and “# 100” match, and the corresponding CIF code “123-45678” is acquired. As a result, “GETRONICS FOODS CO., LTD AKASAKA” in the input data, that is, the lexical phrase is converted into CIF code “123-45678” which is desired data.

  In summary, the apparatus divides a character string such as an input name into basic words, converts and extracts each into a predetermined logical index, and includes an index / corresponding to a CIF code or the like in the combination element. It is determined whether or not the set is included, and when included, a CIF code or the like is derived for the input name or the like.

  Here, an advantage of using the logical index will be described. As described above, by using a logical index, a word originally composed of characters (1 character = 1 byte) is converted into a binary number, and a dictionary search is performed based on this. For example, the word “Shoyama” consisting of 7 characters normally requires 7 bytes to expand the memory as it is. However, when the logical index “207” is applied to this, “207” is expressed as a bit in binary. Then, “11001111” is obtained, which can be expressed by 8 bits (= 1 byte). Therefore, the logical index makes the number of CPU processing units smaller than any word represented by it. By using such a logical index, memory load and CPU load are reduced, and processing speed can be increased.

  The processing blocks described in the processing engine 10 in FIG. 10 and the processing steps in FIG. 13 are the steps 42 and 44 in FIG. 13 being the word recognition block 20 in FIG. 10 and the steps 46 and 48 in FIG. Each corresponds to the name recognition block 22.

  Next, modifications of the above-described apparatus will be described below. The description of the same configuration and operation as the above apparatus will be omitted, and only the different parts will be described. The processing engine 10 expands the basic word dictionary 14 on the main memory 12 before receiving the input data, but does not expand the name dictionary 16 on the main memory 12 in advance.

  Next, the processing engine 10 receives the input data and performs the processing up to steps 44 and 46 in FIG. The processing engine 10 then extracts all of the logical index sets, including the acquired logical index corresponding to the words included in the input data, from the name dictionary 16 and stores them in the main memory 12 as shown at 16a in FIG. Expand memory above.

  Next, the processing engine 10 performs processing similar to step 48 of FIG. Specifically, the processing engine 10 determines that an arbitrary set of the obtained logical indexes “# 107”, “# 106”, “# 104”, and “# 100” corresponding to the words included in the input data is It is searched whether there is a logical index set in the name dictionary 16a expanded in memory shown in FIG. 12, and if they match, “123-45678” is acquired as the CIF code corresponding to the logical index set. As a result, “GETRONICS FOODS CO., LTD AKASAKA” in the input data, that is, the lexical phrase is converted into CIF code “123-45678” which is desired data. In this modified form, the conversion processing speed is slower than that of the above-mentioned apparatus, but the capacity of the main memory 12 may be small.

  In the above-described apparatus, a logical index is defined as a fixed value in advance, the logical index is stored in the basic word dictionary 14 in association with a basic word, and this is expanded in the memory on the main memory 12, so that the key is programmed. Since there is no need to convert the memory address to a memory address that can change each time the memory is loaded, it is not necessary to replace the key relationship between dictionaries with the acquired memory address and generate it. Can be reduced, and the load time can be increased. Further, the analysis process itself can be speeded up by the dictionary analysis composed of the logical index and the logical index set.

  Further, another modification of the above-described apparatus will be described below. In this modification, a plurality of words can be stored in one row of the word column of the basic word dictionary 14. In the above-described apparatus, an analysis process for converting a word into a logical index is performed. Some words to be analyzed have different meanings but have the same meaning in the analysis process. Although different logical indexes may be assigned to these words individually, they have the same semantic content in the analysis process, so even if the same logical index is assigned, no problem occurs in the analysis process, and there is a waste in processing efficiency. It can be omitted. Therefore, a plurality of words having the same semantic content in the analysis processing are stored in one row of the word column of the basic word dictionary 14 although such notation is different. As an example of the storing method, a space may be inserted between words so that each word can be identified. In the above apparatus, any method may be used for storing a plurality of words as long as each word can be identified. For example, “FOODS” and “FOOD” are words having different meanings in the analysis process but different in notation in the analysis process. These words are stored by inserting a space between “FOODS” and “FOOD” as indicated by reference numeral 102 in FIG. Therefore, the same logical index “# 106” is allocated to “FOODS” and “FOOD”.

  When two words “FOODS” and “FOOD” are stored in the word field of the basic word dictionary 14, these words include “FOODS” including two words in the word field as indicated by reference numeral 104. The memory is expanded on the main memory 12 in the format of “FOOD”. Therefore, on the main memory 12, the word “FOOD” is also associated with the same logical index “# 106” as the word “FOODS”.

  In the input data shown in FIG. 13, when “FOOD” is input instead of “FOODS”, “FOOD” is converted into the logical index “# 106” by the same operation as the above-described apparatus, and FIG. “GETRONICS FOOD CO., LTD AKASAKA” is converted to “123-45678” by the same operation as in steps 46 and 48.

  FIG. 14 is a diagram showing an overview for facilitating understanding of a method of assigning the same logical index to a plurality of words having different notations having the same meaning and content in the analysis processing described above. Reference numeral 110 indicates a correspondence relationship between the words and the logical index in the basic word dictionary 14, the word “Syoyama” is different from the word “Shoyama”, and the word “&” is different from the word “and”, but is analyzed. Since the word “Syoyama” is assigned the same logical index “# 207” as the word “Shoyama”, the word “&” is assigned the same logical index “# 206” as the word “and”. It is. Reference numeral 112 indicates the correspondence between the logical index set in the name dictionary 16 and the CIF code. When a lexical word as shown in input 1 is inputted, it is divided into words as shown in division 1, and the word-logical index correspondence relationship indicated by reference number 110 is referred to for each divided word. That is, when a dictionary is applied, it is converted into a logical index shown in dictionary application. On the other hand, when a lexical word as shown in input 2 is input, it is divided into words as shown in division 2, and the correspondence between the word indicated by reference number 110 and the logical index is assigned to each divided word. When referring, that is, applying a dictionary, it is converted to the same logical index as input 1 as shown in Dictionary application. Since the input 1 and the input 2 are converted into the same set of logical indexes, when the correspondence relationship between the logical index set indicated by reference numeral 112 and the CIF code is referred to in the CIF analysis, both the input 1 and the input 2 have the same CIF code. Is converted to

  In the above-described modification in which the same logical index is allocated to a plurality of words having the same meaning and content in the analysis process, the same logical index is allocated to the plurality of words because they have the same meaning and content in the analysis process. By storing the plurality of words in the word column of the basic word dictionary 14 in association with the same logical index, the resources of the storage device such as the main memory and the hard disk are saved, and the processing efficiency is improved. .

  Next, the amount of data that the operator must input when adding words with the same semantic content to the basic word dictionary in the analysis process, although they are different, is reduced as much as possible, reducing errors, and highly reliable A preferred embodiment of the present invention for generating a basic word dictionary database is described below.

First, the case of using the conventional concept of CIF specification without applying the present invention will be described with reference to FIGS.
FIG. 3 shows the contents of the name dictionary 16 and the basic word dictionary 14 of FIG. In this case, it is assumed that the word “&” is not registered in the basic word dictionary 14. FIG. 4 shows a process until input data is converted into a logical index. Assume that input data “THE NAKAMORI TRUST & BANK CORP” is input as indicated by 1 in FIG. In FIG. 4 2, the space is used as a clue to break it down into words. The logical index is obtained by referring to the basic word dictionary 14a on the main memory 212 in 3 of FIG. Since “&” is not registered in the basic word dictionary 14, it remains as it is. As a result, as shown by 4 in FIG. 4, the input data is converted into a logical index except “&”. In the conventional matching process, the logical index “# 501, # 105, # 402, # 101, # 220” of the input data in FIG. 4 is used, and the logical index that cannot be searched even if the name dictionary 16a in FIG. 3 is searched. Since there is “# 301”, the CIF cannot be specified, and analysis NG occurs. An example of this processing flow is shown in part of FIG. In step 300, the basic word dictionary 14 and the name dictionary 16 are expanded in the main memory 212. In step 312 of the input data analysis 310, the input data shown in 1 of FIG. 4 is input, and the input data is divided into words using a space as a clue. Subsequently, in step 314, the divided words are converted into logical indexes corresponding to the words in the basic word dictionary 14a expanded in the memory. In step 322 of the matching process 320, the name dictionary 16a is searched with the logical index of the input data. Next, in step 324, if matching is established between the search results, that is, the logical index set of the input data and the name dictionary 16a, the result is OK and the CIF is specified (step 330). On the other hand, if the matching is not established, the analysis is NG (step 332).

  FIG. 1 shows a preferred embodiment of the apparatus of the present invention which is used in the apparatus described with reference to FIG. 10 and registers words having the same semantic content in the analysis process, although having different notations, in the basic word dictionary. . In FIG. 1, 210 is a processing engine composed of an arbitrary data processing device such as a main frame computer, personal computer, or microprocessor, 212 is a main memory, 14 is a basic word dictionary, and 16 is a name. Each dictionary is shown. The processing engine 210 and the main memory 212 may be the same as or different from the processing engine 10 and the main memory 12 of the apparatus described above. The basic word dictionary 14 and the name dictionary 16 are the same as those of the aforementioned device. Reference numeral 220 denotes an analysis NG logical index set file that stores a logical index that cannot be identified as a CIF and is an analysis NG and an incomplete logical index set including unconverted words (see the logical index set 400 shown in 4 of FIG. 4). It is. The basic word dictionary 14, the name dictionary 16, and the analysis NG logical index set file 220 are stored in a hard disk (not shown) such as a magnetic disk, but are not limited to this, and are in any other format. It can be stored in a storage device. Reference numeral 222 denotes an input device such as a keyboard for inputting data and the like, and 224 denotes a display device such as a normal liquid crystal display for displaying data and the like. The data processing device that functions as the processing engine 210, the main memory 212, the basic word dictionary 14, the name dictionary 16, and the hard disk that stores the analysis NG logical index set file 220, the input device 222, and the display device 224 They are connected to each other by a normal data bus or the like (not shown). Note that the components and functions of the apparatus shown in FIG. 1 can be expanded and incorporated in the apparatus shown in FIG.

  The processing engine 210 executes steps 300 to 332 of the processing flow of FIG. In the case of analysis NG, the processing engine 210 stores an incomplete logical index set including unconverted words as indicated by reference numeral 400 in FIG. 4 in the analysis NG logical index set file 220. Note that these functions of the processing engine 210 may be incorporated into the processing engine 10 of the above-described apparatus. Further, the processing engine 10 of the above-described apparatus executes steps 300 to 332, and in the case of analysis NG, an incomplete logical index set including an unconverted word as indicated by reference numeral 400 in FIG. Processing to store in the analysis NG logical index set file 220 may be performed.

Next, a processing flow to be registered in the basic word dictionary 14 will be described with reference to FIGS.
In step 342 of the input data acquisition process 340 in FIG. 9, the operator instructs the input device 222 to display the incomplete logical index set stored in the analysis NG logical index set file 220 on the display device 224. input. Next, in response to the command, the processing engine 210 displays the incomplete logical index set stored in the analysis NG logical index set file 220 on the display device 224. The display method of the displayed incomplete logical index set may be in a list format, limited to a specific one, and is arbitrary according to the usage pattern. The incomplete logical index set 402 described in 1 of FIG. 5 is a logical index set that is analyzed NG because the CIF cannot be specified in the above example.

  In step 344 of FIG. 9, the processing engine 210 obtains input data when the operator makes a correction input to the incomplete logical index set 402 displayed on the display device 224. Specifically, as shown in 2 of FIG. 5, the operator inputs the CIF code “9130,000” (see 404 of FIG. 5) estimated from the displayed incomplete logical index set 402 from the input device 222.

  In step 352 of the basic word dictionary registration processing 340 in FIG. 9, the processing engine 210 searches the name dictionary 16 using the corrected CIF code “9130,000” as a key, and acquires a logical index set from the name dictionary 16. . In the example of FIG. 5, the row of the CIF code “9130,000” of the name dictionary 16 described in FIG. 3 is selected, and the logical index set “# 501, # 105, # 402, # 301, # 101, # 220” is selected. (Reference number 406 in 2 of FIG. 5).

  Subsequently, the processing engine 210 performs a registered word determination process. Specifically, in step 354 of FIG. 9, the processing engine 210 determines that the logical index set “# 501, # 105, # 402, &, # 101, ## including the unacquirable word“ & ”that has been analyzed NG. 220 ”(408 in 3 of FIG. 5) and the logical index set“ # 501, # 105, # 402, # 301, # 101, # 220 ”of the name dictionary 16 acquired in 2 of FIG. 5 (3 in FIG. 5). 410), and the logical indexes excluding those that are the same between the two, that is, the logical indexes of the differences are specified. As indicated by 3 in FIG. 5, the logical indexes “# 501”, “# 105”, “# 402”, “# 101”, and “# 220” are the logical index of the data that has been analyzed NG and the name dictionary. Since it is the same logical index which exists in both logical indexes, it excludes. The word “&” (see 412 in 3 of FIG. 5) and the logical index “# 301” (see 414 of 3 in FIG. 5) of the name dictionary 16 are acquired as an analysis NG that is excluded from the exclusion. It should be noted that it is preferable that the number of the analysis NG that is excluded from the exclusion, that is, the difference word is one.

  Subsequently, in step 356 of FIG. 9, the processing engine 210 performs a dictionary registration process (see also 4 in FIG. 5). In the example of FIG. 5, the word “&” of analysis NG acquired in the registered word determination process is registered as a synonym in the basic word dictionary 14. Specifically, as indicated by 4 in FIG. 4 and reference numeral 416 in FIG. 6, the identified difference word “&” is registered in the “AND” field registered in the basic word dictionary 14. A column in which “&” is added to the registered “AND” is as indicated by reference numeral 418 in FIG. 6 when the basic word dictionary 14 is expanded in memory.

  Next, with respect to the basic word dictionary 14 after registration in FIG. 6 (the name dictionary 16 is the same as shown in FIG. 3 without change), the data “THE NAKAMORI of the analysis NG in FIG. 4 described above is used. A case where “TRUST & BANK CORP” is input will be described with reference to FIGS. 7 and 8.

7 shows a state in which the basic word dictionary 14 after registration in FIG. 6 and the same name dictionary 16 as in FIG. 3 are expanded in the main memory 212 or 12 shown in FIG.
The processing engine 210 (or 10) converts the input data into a logical index using the memory expanded basic word dictionary 14a shown in FIG. 7, as shown in 1-4 of FIG. At this time, as shown in FIG. 6 and FIG. 7, “AND” and “&” are registered in the basic word dictionary 14 so that the logical index “# 301” can be acquired. All data words are converted into logical indexes, and the converted logical index set is “# 501, # 105, # 402, # 301, # 101, # 220” (see 420 in 4 of FIG. 8). When the name dictionary 16a of FIG. 7 is searched using this logical index set, logical index sets “# 501, # 105, # 402, # 301, # 101, # 220” registered in the name dictionary 16a are obtained. Since the search is possible, the CIF “9130,000” can be specified.

  As a result, when registering the input data and the word in the name dictionary 16 as a synonym, a state where the difference can be specified as 1: 1 is registered as a synonym. Specific factors can be increased.

  The present invention incorporates the functions of the processing engine 210 into the processing engine 10 and does not have the analysis NG logical index set file 220, and the processing engine 10 continuously performs steps 300 to 356 in the processing flow of FIG. You may make it perform to. In that case, the processing engine 10 displays the analysis NG in step 332 on the display device when the search result in step 324 becomes analysis NG, and simultaneously displays the incomplete logical index set in step 342 on the display device. . Subsequent processing of the processing engine 10 is the same as described above with reference to FIG.

FIG. 1 is used for a device that converts a customer's customer information (CIF) analysis process using a computer to a payment instruction message received at a bank, etc. 1 shows a preferred embodiment of an apparatus of the present invention for registering words having the same semantic content in processing in a basic word dictionary. FIG. 2 shows a state in which basic words registered in advance in the basic word dictionary 14 are expanded in the main memory 12. FIG. 3 shows a state where a logical index set registered in advance in the name dictionary 16 is expanded in the main memory 12. FIG. 4 shows an example in which input data is converted into a logical index and becomes analysis NG. FIG. 5 shows a process of registering words having the same semantic contents in the analysis process in the basic word dictionary, although the notations are different. FIG. 6 shows the basic word dictionary 14 after the words having the same meaning and content in the analysis processing are registered in the basic word dictionary, although the notations are different, together with the state of memory expansion. FIG. 7 shows the expanded state of the basic word dictionary 14 and the name dictionary 16 after registration shown in FIG. FIG. 8 shows an example in which the input data can be analyzed using the basic word dictionary 14 after the words having the same meaning and content in the analysis processing are registered in the basic word dictionary, although the notations are different. FIG. 9 shows a processing flow for registering words having the same semantic content in the analysis processing, although the notations are different, in the basic word dictionary. FIG. 10 is a diagram showing a basic configuration of an example of an apparatus that converts a lexical phrase used for performing a customer customer information (CIF) analysis process using a computer to a payment instruction message received by a bank or the like into data. is there. FIG. 11 shows a state in which basic words registered in advance in the basic word dictionary 14 are expanded in the main memory 12. FIG. 12 shows a state in which a logical index set registered in advance in the name dictionary 16 is expanded on the main memory 12. FIG. 13 is a diagram for explaining a process in which data input to the conversion device shown in FIG. 10 is converted. FIG. 14 is a diagram showing an overview for facilitating understanding of the invention in which the same logical index is assigned to a plurality of differently expressed words having the same meaning and content in the analysis process.

Explanation of symbols

10, 210 Processing engine 12, 212 Main memory 14 Basic word dictionary 16 Name dictionary 220 Analysis NG logical index set file 222 Input device 224 Display device

Claims (9)

An apparatus for converting a lexical word including a plurality of words into another data including one piece of information specified by the plurality of words,
At least a word field that stores words in advance and a logical index field that stores a logical index that represents the words stored in the word field, and each of the plurality of words represents each word. Are registered in advance in association with the logical index to be processed, and the word column has the same logical index attached to a plurality of words having the same meaning and different notation in the analysis processing for converting the word into the logical index. A basic word dictionary configured to be capable of storing the plurality of words,
A set of logical indexes corresponding to a set of words that are stored in advance in the storage means and registered in the basic word dictionary is associated with one piece of information specified by the set of words and registered in advance. Name dictionary,
Associating a plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis processing with the same logical index assigned to the plurality of words having different notations The memory is expanded on the memory, and the set of logical indexes registered in the name dictionary is associated with one piece of information specified by the set of words corresponding to each set of logical indexes, and the memory is expanded on the memory. , Decomposing a lexical phrase including a plurality of words into words, referring to a basic word dictionary expanded in memory on the memory, obtaining a logical index corresponding to the decomposed word, and extracting the logic expanded in the memory Any one of the logical indexes composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed word from the set of indexes. By selecting a set of logical indexes that matches the set of indexes and using one piece of information specified by the set of selected logical indexes, the lexical phrase including the plurality of words includes the one piece of information specified. When converting to another data and acquiring a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word is determined among the decomposed words A first processing engine for storing in a parsing NG logical index set file stored in the storage means together with the logical index of the word for which the logical index could be obtained;
An input means for inputting information including data;
Display means for displaying information including data;
In response to a search command input from the input means, search the word that could not obtain at least the corresponding logical index from the analysis NG logical index set file and the logical index stored together with the word and display The name dictionary using, as a key, one piece of information estimated from a word that is displayed on the input means and is input from the input means and for which the logical index displayed on the display means cannot be obtained and the logical index stored together with the word To obtain a set of logical indexes corresponding to the one piece of estimated information, a word for which the logical index could not be obtained, a logical index stored with the word, and a logical index obtained from the name dictionary. Identify the difference words and logical index from the set of indexes A second processing engine of registering words in the difference as a synonym to the word column within said basic word dictionary corresponding to the logical indexes of the difference,
A device characterized by comprising: An apparatus for converting a lexical word including a plurality of words into another data including one piece of information specified by the plurality of words,
At least a word field that stores words in advance and a logical index field that stores a logical index that represents the words stored in the word field, and each of the plurality of words represents each word. Are registered in advance in association with the logical index to be processed, and the word column has the same logical index attached to a plurality of words having the same meaning and different notation in the analysis processing for converting the word into the logical index. A basic word dictionary configured to be capable of storing the plurality of words,
A set of logical indexes corresponding to a set of words that are stored in advance in the storage means and registered in the basic word dictionary is associated with one piece of information specified by the set of words and registered in advance. Name dictionary,
An input means for inputting information including data;
Display means for displaying information including data;
Associating a plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis processing with the same logical index assigned to the plurality of words having different notations The memory is expanded on the memory, and the set of logical indexes registered in the name dictionary is associated with one piece of information specified by the set of words corresponding to each set of logical indexes, and the memory is expanded on the memory. , Decomposing a lexical phrase including a plurality of words into words, referring to a basic word dictionary expanded in memory on the memory, obtaining a logical index corresponding to the decomposed word, and extracting the logic expanded in the memory Any one of the logical indexes composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed word from the set of indexes. By selecting a set of logical indexes that matches the set of indexes and using one piece of information specified by the set of selected logical indexes, the lexical phrase including the plurality of words includes the one piece of information specified. When converting to another data and acquiring a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word is determined among the decomposed words The logical index of the word that has been obtained with the logical index is displayed on the display unit together with the logical index of the word that has been obtained, and is input from the input unit. The name is used as a key with one piece of information estimated from the logical index of the word for which A dictionary is searched to obtain a set of logical indexes corresponding to the one piece of estimated information, and a logical index of a word for which a logical index can be obtained from the word that could not obtain the logical index and the decomposed word And the difference word and the logical index with the set of logical indexes acquired from the name dictionary, and the difference word as a synonym in the word field in the basic word dictionary corresponding to the difference logical index A processing engine to register,
A device characterized by comprising: A method of converting a lexical word including a plurality of words into another data including one piece of information specified by the plurality of words,
At least a word field for storing words and a logical index field for storing a logical index representing the words stored in the word field, and each of the plurality of words is associated with a logical index representing each word in advance. The plurality of words are registered, and the word column has the same logical index attached to a plurality of words having the same semantic content and different notation in the analysis processing for converting the word into a logical index. Storing a basic word dictionary configured to be stored in a storage means;
A logical dictionary set corresponding to a set of words obtained by combining words registered in the basic word dictionary is stored as a name dictionary registered in advance in association with one piece of information specified by the set of words. Memorizing step;
Associating a plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis processing with the same logical index assigned to the plurality of words having different notations Expanding the memory on the memory; and
Expanding a set of logical indexes registered in the name dictionary in the memory in association with one piece of information specified by the set of words corresponding to each set of logical indexes;
Decomposing a lexical word to be converted containing multiple words into words;
Obtaining a logical index corresponding to the decomposed word by referring to a basic word dictionary expanded in memory on the memory;
A logical index set that matches any logical index set composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed words is selected from the memory-expanded logical index sets; Converting a lexical phrase including the plurality of words into another data including the specified one information by using one information specified by the selected set of logical indexes;
When acquiring a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word is identified as a word of the word that has acquired the logical index among the decomposed words. Storing together with the logical index in an analysis NG logical index set file stored in the storage means;
In response to a search command input from input means for inputting information including data, at least the corresponding logical index could not be obtained from the analysis NG logical index set file and stored together with the word Retrieving a logical index and displaying it on a display means for displaying information including data;
The name dictionary is searched by using, as a key, a word input from the input unit and the information that is estimated from the logical index stored together with the word that could not be obtained from the logical index displayed on the display unit, Obtaining a set of logical indexes corresponding to the one piece of estimated information;
A difference word and a logical index between a word for which the logical index could not be obtained and a logical index stored together with the word and a set of logical indexes obtained from the name dictionary are specified, and the difference corresponds to the logical index Registering the difference word as a synonym in the word field in the basic word dictionary;
A method characterized by comprising : A method of converting a lexical word including a plurality of words into another data including one piece of information specified by the plurality of words,
At least a word field for storing words and a logical index field for storing a logical index representing the words stored in the word field, and each of the plurality of words is associated with a logical index representing each word in advance. The plurality of words are registered, and the word column has the same logical index attached to a plurality of words having the same semantic content and different notation in the analysis processing for converting the word into a logical index. Storing a basic word dictionary configured to be stored in a storage means;
A logical dictionary set corresponding to a set of words obtained by combining words registered in the basic word dictionary is stored as a name dictionary registered in advance in association with one piece of information specified by the set of words. Memorizing step;
Associating a plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis processing with the same logical index assigned to the plurality of words having different notations Expanding the memory on the memory; and
Expanding a set of logical indexes registered in the name dictionary in the memory in association with one piece of information specified by the set of words corresponding to each set of logical indexes;
Decomposing a lexical word to be converted containing multiple words into words;
Obtaining a logical index corresponding to the decomposed word by referring to a basic word dictionary expanded in memory on the memory;
A logical index set that matches any logical index set composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed words is selected from the memory-expanded logical index sets; Converting a lexical phrase including the plurality of words into another data including the specified one information by using one information specified by the selected set of logical indexes;
When acquiring a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word is identified as a word of the word that has acquired the logical index among the decomposed words. Displaying together with the logical index on a display means for displaying information including data;
Inferred from the logical index of the word that was able to acquire the logical index among the words that could not be acquired the logical index displayed on the display means and the decomposed words that were input from the input means for inputting information including data Searching the name dictionary using one piece of information as a key to obtain a set of logical indexes corresponding to the one piece of estimated information;
Identifying a word and a logical index of a difference between a logical index of a word from which the logical index could not be obtained and a logical index of a word from which the logical index could be obtained, and a set of logical indexes obtained from the name dictionary; Registering the difference word as a synonym in the word field in the basic word dictionary corresponding to the difference logical index;
A method characterized by comprising : At least a word field for storing words and a logical index field for storing a logical index representing the words stored in the word field, and each of the plurality of words is associated with a logical index representing each word in advance. The plurality of words are registered, and the word column has the same logical index attached to a plurality of words having the same semantic content and different notation in the analysis processing for converting the word into a logical index. A set of logical indexes corresponding to a set of words that is a combination of words registered in the basic word dictionary, and one piece of information specified by the set of words, Storage means for storing a name dictionary registered in advance in association;
An input means for inputting information including data;
Display means for displaying information including data;
A program for causing a computer including a plurality of words to function as a device that converts a lexical phrase including a plurality of words into another data including one piece of information specified by the plurality of words,
The computer,
Associating a plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis processing with the same logical index assigned to the plurality of words having different notations The memory is expanded on the memory, and the set of logical indexes registered in the name dictionary is associated with one piece of information specified by the set of words corresponding to each set of logical indexes, and the memory is expanded on the memory. , Decomposing a lexical phrase including a plurality of words into words, referring to a basic word dictionary expanded in memory on the memory, obtaining a logical index corresponding to the decomposed word, and extracting the logic expanded in the memory Any one of the logical indexes composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed word from the set of indexes. By selecting a set of logical indexes that matches the set of indexes and using one piece of information specified by the set of selected logical indexes, the lexical phrase including the plurality of words includes the one piece of information specified. When converting to another data and acquiring a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word is determined among the decomposed words A first processing engine that stores a logical index of a word for which a logical index has been acquired in an analysis NG logical index set file stored in the storage means;
And, in response to the search command input from the input means, at least the corresponding logical index from the analysis NG logical index set file and the logical index stored together with the word are searched. The information displayed on the display means and input from the input means, the logical index displayed on the display means for which the logical index was not obtained, and one information estimated from the logical index stored together with the word as a key. The name dictionary is searched to obtain a set of logical indexes corresponding to the one piece of estimated information, and the logical index stored together with the words for which the logical index cannot be obtained and the words are obtained from the name dictionary. The difference word and logical index from the set of logical index Constant, and a second processing engine of registering words in the difference in the word column within said basic word dictionary corresponding to the logical indexes of the difference as a synonym
A program characterized by functioning as At least a word field for storing words and a logical index field for storing a logical index representing the words stored in the word field, and each of the plurality of words is associated with a logical index representing each word in advance. The plurality of words are registered, and the word column has the same logical index attached to a plurality of words having the same semantic content and different notation in the analysis processing for converting the word into a logical index. A set of logical indexes corresponding to a set of words that is a combination of words registered in the basic word dictionary, and one piece of information specified by the set of words, Storage means for storing a name dictionary registered in advance in association;
An input means for inputting information including data;
Display means for displaying information including data;
A program for causing a computer including a plurality of words to function as a device that converts a lexical phrase including a plurality of words into another data including one piece of information specified by the plurality of words,
The computer,
Associating a plurality of words that are registered in the basic word dictionary and have the same semantic content and different notation in the analysis processing with the same logical index assigned to the plurality of words having different notations The memory is expanded on the memory, and the set of logical indexes registered in the name dictionary is associated with one piece of information specified by the set of words corresponding to each set of logical indexes, and the memory is expanded on the memory. , Decomposing a lexical phrase including a plurality of words into words, referring to a basic word dictionary expanded in memory on the memory, obtaining a logical index corresponding to the decomposed word, and extracting the logic expanded in the memory Any one of the logical indexes composed of a plurality of logical indexes among the logical indexes corresponding to the decomposed word from the set of indexes. By selecting a set of logical indexes that matches the set of indexes and using one piece of information specified by the set of selected logical indexes, the lexical phrase including the plurality of words includes the one piece of information specified. When converting to another data and acquiring a logical index corresponding to the decomposed word, if there is a word for which a logical index cannot be acquired, the word is specified, and the specified word is determined among the decomposed words The logical index of the word that has been obtained with the logical index is displayed on the display unit together with the logical index of the word that has been obtained, and is input from the input unit. The name is used as a key with one piece of information estimated from the logical index of the word for which A dictionary is searched to obtain a set of logical indexes corresponding to the one piece of estimated information, and a logical index of a word for which a logical index can be obtained from the word that could not obtain the logical index and the decomposed word And the difference word and the logical index with the set of logical indexes acquired from the name dictionary, and the difference word as a synonym in the word field in the basic word dictionary corresponding to the difference logical index Processing engine to register
A program characterized by functioning as The apparatus according to claim 1, wherein the difference word is one. 5. The method according to claim 3, wherein the difference word is one. 7. The program according to claim 5, wherein the difference word is one.

Images