JP2000207395A - Device and method for analyzing japanese language and storage medium recording japanese language analyzing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device and method for analyzing Japanese language by which a dictionary storing area can be reduced and the processing efficiency can be improved and a storage medium on which a Japanese language analyzing program is recorded. SOLUTION: In a Japanese language analyzing device having two dictionaries for KANA/KANJI conversion and for Japanese language analysis, the KANA/ KANJI conversion index storing section 3 of a KANA/KANJI-converting dictionary section 2 and the Japanese language analyzing dictionary index storing section 7 of a Japanese language analyzing dictionary section 6 are integrated by means of an index information merging section 10. Therefore, the used amount of memories is significantly reduced, because the index information of the two dictionaries is commonly used.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is incorporated in a machine translation device for translating Japanese into another language, a Japanese proofreading support device for pointing out and correcting an input error in input Japanese, and the like. The present invention relates to a Japanese language analyzing device, a Japanese language analyzing method, and a recording medium on which a Japanese language analyzing program is recorded.

[0002]

2. Description of the Related Art In recent years, with the spread of personal computers and the development of information and communication networks, machine translation devices for translating Japanese into other languages and correction of Japanese input errors have been corrected.
Japanese-language processing devices such as the Japanese-language proofreading support devices that are pointed out have rapidly spread. It is necessary for Japanese language processors to perform the so-called Japanese analysis, which grasps the syntax and semantic structure of the Japanese sentence to be processed, and all Japanese processors are implemented with a built-in Japanese language analyzer. I have.

[0003] The Japanese language analyzing apparatus includes: (1) a morphological analysis process for dividing a Japanese sentence given as a character string into words while referring to a word dictionary; and (2) a relation between words based on the divided words. This is realized by sequentially performing syntax and semantic analysis processing for performing a receiving analysis and a semantic analysis and determining a unique syntax and semantic structure.

In the morphological analysis process, all words starting from an arbitrary position in an input sentence are searched from a word dictionary and passed to a syntax / semantic analysis process. Then, in the syntactic and semantic analysis processing, a syntactic and semantic structure that satisfies grammatical constraints such as part of speech and semantic constraints from this word candidate, that is, A structure such as that shown in FIG. 9 showing whether the object plays a role is shown as a tree structure.

At this time, in the syntax / semantic analysis processing, the part of speech and the meaning of the word (the meaning of the word) are determined. In the Japanese analysis, the dependency relation of each word is represented by the upper and lower relations of the tree structure, and the syntactic and semantic structure as shown in FIG. The purpose of.

However, even if the syntax is exactly the same, there are cases where the syntax and semantic structure differ depending on the constituent words of the sentence. For example, the following statement is used.

(1) He has black eyes.

(2) He likes her.

[0009] Although the above two sentences have completely equivalent syntaxes of "~ wa predicate", their syntax and semantic structure have completely different structures as shown in Figs. ing. This is because the meanings of the constituent words of the sentence are different. In other words, if the meaning of the word, that is, the meaning of the word, is not clearly defined, the syntactic / semantic structure that is the ultimate goal of the syntactic / semantic analysis cannot be output. That is, this semantic determination is an important issue in Japanese language analysis processing, and greatly affects the analysis accuracy.

Here, a description will be given of the semantic determination in the conventional Japanese analysis.

Conventionally, in order to determine the part of speech and the meaning of a word, a Japanese parsing apparatus discards low-probability ones using a method of pruning and selects the most probable one from the remaining ones. To make the final solution.

For example, in the input sentence "His head is wonderful", the word "head" has the interpretation of a noun and a suffix. In the syntactic and semantic analysis processing, first, the interpretation of the suffix is discarded based on the grammatical constraint that “the suffix connects immediately after the independent word”. Next, the meaning of a noun is determined. There are various meanings of "head", such as "body part", "tip of object", "brain", and "head of craftsman". In order to finally determine a unique meaning from this meaning, the current Japanese analysis processing is performed based on whether or not the semantic constraint of the declinable word “great” is met. Usually, semantic restrictions on cases that can be taken by the verbal are described as dictionary information on the verbal. In the example, the semantic constraint that "a person or a brain can come as a constraint on a ga case" is described for the verb "excellent". Meanings that meet the above constraints from the plural meanings of the case "head",
In other words, pick out the meaning. In the example, "the brain" and "the artisan's leadership" conform to the semantic constraints described in the adjective "great."
Will be selected. The above-described constraints are repeatedly applied until the meanings of all words are finally determined uniquely.

As described above, it is necessary to determine the meaning of a word in the Japanese analysis. However, as shown in the above example, when determining the meaning of a word, the meaning of the word obtained as a result of searching a dictionary is determined. It is performed for all parts of speech and meanings. However, at the time of dictionary search, since the dictionary search is performed only for the character strings included in the input sentence, that is, for the word face of the word, and all words registered in the dictionary are taken out, there is a possibility that the number of word meaning determination candidates may be increased at random. There is. In the semantic determination processing, it is necessary to perform the semantic determination while trying as many combinations as possible for this semantic determination candidate, and as a result, a problem occurs in that the efficiency and accuracy of the meaning hair determination processing are reduced. . In order to solve this problem, unnecessary interpretations (senses) should be made in advance of the Japanese analysis using kana information input through gravity means such as a keyboard and a voice input device attached to the information terminal device. There is a method of discarding and reducing the number of semantic determination candidates. For example, in the above example, it is assumed that “He is wonderful” is input through the keyboard, and the sentence to be analyzed by the Japanese language “He is wonderful” is converted through the kana-kanji conversion device.
It is assumed that the reading for "head" is stored as "head". In this case, when performing a dictionary search for “head” in a dictionary search at the time of Japanese analysis, the search is performed while referring to the pronunciation at the same time, so that “head (read: kashira)”, “head (read: Words) can be excluded from the dictionary search results. As a result, it is possible to squeeze the target candidates in the meaning determination, and to improve the analysis accuracy and the analysis efficiency. That is, by combining the kana-kanji conversion device and the Japanese analysis device, the analysis accuracy and efficiency can be improved.

[0014]

At present, kana-kanji conversion devices installed in many information devices also belong to the category of so-called Japanese language processing devices, and have a parsing unit and a parsing dictionary inside. I have. However, when a new Japanese-language processing device, for example, a machine translation device, is incorporated into information equipment,
The processing target is different from the kana-kanji conversion device and the kana-kanji character string, and the machine translation device is different from the kana-kanji mixed character string, but the same word vocabulary information (for example, as vocabulary information)
Exists. Nevertheless, a dictionary for kana-kanji conversion,
Since there is a dictionary for machine translation, the storage area of the dictionary is wasted.

According to the present dictionary search method, index information, which is address information to dictionary data, is stored in a memory, and the dictionary data is read from the index as needed (every time a dictionary search request is made). Is realized by the method of extracting data into the memory. Therefore, when there are a plurality of dictionaries, not only is a large storage area for this index necessary, but also in each application program, the data retrieved from the memory This is because there may be information duplication.

Furthermore, in order to perform the current Kana-Kanji conversion and Japanese analysis, it is necessary to perform a dictionary search and a syntax analysis, respectively, which causes a reduction in processing efficiency.

According to the present invention, when kana-kanji information which is an input to a kana-kanji conversion apparatus is used for the semantic determination processing of Japanese-language analysis, vocabulary data of a dictionary necessary for performing kana-kanji conversion and Japanese analysis. The purpose of the present invention is to provide a Japanese-language analyzer, a Japanese-language analysis method, and a recording medium on which a Japanese-language analysis program is recorded, which can reduce the amount of memory used as an index to a computer and save valuable resources in information equipment I do.

[0018]

SUMMARY OF THE INVENTION In order to solve this problem, the present invention provides an input section for inputting kana characters, instructing the start of kana-kanji conversion and Japanese syntax analysis, a kanji corresponding to a kana character, and A kana-kanji conversion dictionary that stores part-of-speech and syntax information, a kana-kanji conversion dictionary index storage that stores the vocabulary information of words stored in the kana-kanji conversion dictionary, and a kana-kanji conversion dictionary index storage Based on the vocabulary information of the word requested to be obtained from the kana-kanji conversion dictionary, the kana-kanji information acquisition unit and the word kana character input based on the word vocabulary information obtained from the kana-kanji information acquisition unit. A kana-kanji conversion unit for converting to a kanji code,
A Japanese analysis dictionary unit that stores reading and pronunciation, part-of-speech, grammar information, and semantic information for words, and a Japanese analysis dictionary index storage unit that stores the storage location of vocabulary information of words stored in the Japanese analysis dictionary unit , Japanese lexical information acquisition unit that extracts vocabulary information of words requested to be acquired from the Japanese analytic dictionary unit via the Japanese analytic dictionary index storage unit, and word lexical information obtained from the kana / kanji information acquisition unit Based on the Kana-Kanji conversion unit, a Japanese-language analysis unit that performs Japanese morphological analysis, syntactic and semantic analysis of Kana-Kanji mixed sentences, a Kana-Kanji conversion dictionary index storage unit, and a Japanese analysis dictionary index storage unit The index information integration unit that integrates the index information in, the index of the kana-kanji conversion dictionary, the index of the Japanese analysis dictionary, the kana-kanji conversion, An analysis information storage unit that stores time analysis of Japanese analysis, is obtained by a display unit for displaying the Japanese analysis results, the structure having.

Thus, when the kana information input to the kana-kanji conversion device is used for the semantic determination processing of the Japanese analysis, the kana-kanji conversion and the vocabulary data of the dictionary necessary for the Japanese analysis are performed. It is possible to realize a Japanese language analyzer that can reduce the amount of memory used as an index and save valuable resources in information equipment.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The invention according to claim 1 of the present invention provides an input unit for inputting a kana character, a kana-kanji conversion, and a start instruction for Japanese syntax analysis, a kanji corresponding to a kana character, and its part of speech. A kana-kanji conversion dictionary unit that stores word and syntax information, a kana-kanji conversion dictionary index storage unit that stores the storage location of the vocabulary information of the words stored in the kana-kanji conversion dictionary unit, and a kana-kanji conversion dictionary index storage unit. hand,
Kana-kanji information acquisition unit that retrieves the vocabulary information of the requested word from the kana-kanji conversion dictionary unit, and converts the input kana character to kanji code based on the word vocabulary information obtained from the kana-kanji information acquisition unit. A kana-kanji conversion unit for conversion, a Japanese analysis dictionary unit that stores reading and pronunciation, part-of-speech, grammar information, and semantic information for words, and a storage location for vocabulary information of words stored in the Japanese analysis dictionary unit A Japanese analysis dictionary index storage unit, a Japanese analysis information acquisition unit that retrieves the vocabulary information of the requested words from the Japanese analysis dictionary unit via the Japanese analysis dictionary index storage unit, and a kana kanji information acquisition Based on the word vocabulary information obtained from the Kana-Kanji conversion part, the Japanese analysis part that performs Japanese morphological analysis, syntax and semantic analysis of the Kana-Kanji mixed sentence obtained from the Kana-Kanji conversion part, and the Kana-Kanji conversion dictionary Index storage unit, index information integration unit that integrates index information in the Japanese analysis dictionary index storage unit, index of the kana-kanji conversion dictionary unit, index of the Japanese analysis dictionary unit, kana-kanji conversion, and Japanese analysis And a display unit for displaying the results of the Japanese language analysis. When the device is incorporated and the kana information input to the kana-kanji conversion device is used for the semantic determination processing of the Japanese analysis, it has the effect of reducing the amount of memory used.

According to a second aspect of the present invention, in the first aspect of the invention, a kana character is inputted by an input unit, and a kana-kanji conversion unit converts a kana-kanji character string obtained by converting a kana character into a kanji. When passing to the Japanese analysis unit, the kana-kanji information acquisition unit also passes the vocabulary information acquired from the kana-kanji dictionary unit at the same time, and when the Japanese analysis unit searches the Japanese analysis dictionary unit,
Judge whether or not there is overlapping information between the vocabulary information of the word passed from the Kana-Kanji conversion unit, and if there is an overlap, the Japanese analysis information acquisition unit returns from the Japanese analysis dictionary unit This is a configuration that removes duplication from the information to be acquired and acquires only the difference information of the vocabulary information. When using kana information, which is input to the conversion device, for the semantic determination processing of Japanese analysis, it is possible to suppress an increase in memory usage when extracting dictionary data into a memory, and to reduce dictionary data acquisition time. Has an action.

According to a third aspect of the present invention, in the first aspect of the invention, a kana kanji character string is inputted by an input unit, and a kana kanji conversion unit converts the kana kanji character string into a kanji character. When the kana-kanji conversion unit converts the kana character string to kanji when passing it to the Japanese analysis unit, it also passes the analysis result that was performed at the same time, and the Japanese analysis unit uses this analysis information and the kana-kanji conversion unit outputs It is designed to analyze syntax and semantics of the sentence mixed with kana-kanji characters, and the Japanese parser reuses the syntax analysis result passed from kana-kanji conversion and does not perform extra syntax analysis and semantic analysis. Thus, the processing time of the Japanese language analyzer can be reduced.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

Embodiment 1 Hereinafter, a first embodiment of the present invention will be described.

FIG. 1 is a functional block diagram of the Japanese language analyzing apparatus according to the first embodiment of the present invention, and shows the configuration of functional units.

In FIG. 1, reference numeral 1 denotes an input unit for inputting kana characters and instructing the start of kana-kanji conversion and Japanese syntax analysis. Reference numeral 2 denotes a kana-kanji conversion dictionary storing the kanji corresponding to the kana character and its part of speech and syntax information. Reference numeral 3 denotes a kana-kanji conversion dictionary index storage unit that stores the storage locations of the vocabulary information of the words stored in the kana-kanji conversion dictionary unit 2. Reference numeral 4 denotes a kana-kanji information acquisition unit for extracting a word requested to be acquired from the kana-kanji conversion dictionary unit 2 via the kana-kanji conversion dictionary index storage unit 3. Reference numeral 5 denotes a kana-kanji conversion unit that converts input kana characters into kanji codes based on the word vocabulary information obtained from the kana-kanji information acquisition unit 4.

Reference numeral 6 denotes a Japanese analysis dictionary unit which stores readings and pronunciations of words, parts of speech, grammar information, meaning information, and the like. Reference numeral 7 denotes a Japanese analysis dictionary index storage unit that stores the storage locations of the vocabulary information of the words stored in the Japanese analysis dictionary unit 6. 8 is a Japanese analysis dictionary index storage unit 7
Is a Japanese-language analysis information acquisition unit that retrieves a word requested to be acquired from the Japanese-language analysis dictionary unit 6 via the. 9 is based on the word vocabulary information obtained from the Japanese analysis information acquisition unit 8,
This is a Japanese analysis unit that performs Japanese morphological analysis and syntax / semantic analysis on the sentence mixed with kana / kanji obtained from the kana / kanji conversion unit 5.

Reference numeral 10 denotes an index information merge unit for merging the index information in the kana-kanji conversion dictionary index storage unit 3 and the index information in the Japanese analysis dictionary index storage unit 7. Reference numeral 11 denotes an analysis information storage unit that stores an index of the kana-kanji conversion dictionary unit 2, an index of the Japanese analysis unit 9, and analysis information at the time of kana-kanji conversion and Japanese analysis. Reference numeral 12 denotes a display unit for displaying a Japanese analysis result, and reference numeral 13 denotes a control unit for controlling the operation of each unit described above.

FIG. 2 is a circuit block diagram of the registration device and the translation device according to the first embodiment of the present invention, and shows a hardware configuration.

In FIG. 2, reference numeral 14 denotes an input device such as a keyboard or a pointing device. Reference numeral 15 denotes a display device such as a cathode ray tube display (CRT). 1
Reference numeral 6 denotes a central processing unit (CPU) that controls the apparatus by executing various programs. Reference numeral 17 denotes a random access memory (RAM) which is a memory to which data can be written. Reference numeral 18 denotes a read-only memory (ROM) which is a data read-only memory.

Reference numeral 19 denotes a recording medium for recording data such as a CD-ROM, and 20 denotes a recording medium such as a CD-ROM drive.
A reading device for reading data from the recording medium 19, 21 is a data bus.

Here, the correspondence between the functional means of FIG. 1 and the hardware of FIG. 2 will be described.

1 and 2, the input unit 1 is realized by the input device 14, the display unit 12 and the display device 15, and the analysis information storage unit 11 is realized by the RAM 18.

The kana-kanji conversion dictionary unit 2, the kana-kanji conversion dictionary index storage unit 3, the Japanese analysis dictionary unit 6, and the Japanese analysis dictionary index storage unit 7 are a RAM 17 and a ROM 1.
8 is realized.

The kana-kanji information acquisition unit 4, the kana-kanji conversion unit 5, the Japanese analysis information acquisition unit 8, the Japanese analysis unit 9, the index information merge unit 10, and the control unit 13
Is realized by executing various programs stored in the ROM 18 while exchanging data with the RAM 17 and the ROM 18.

Note that, in the present embodiment, the CPU 16
Although the form in which various programs stored in the OM 17 are executed is shown, the programs may be recorded on the recording medium 19.

In this case, the CPU 16 develops the program data read from the recording medium 19 via the reading device 20 into the RAM 17 and executes it. With such an embodiment, the present invention can be easily realized by a general-purpose computer.

The operation of the Japanese language analyzer configured as described above will be described below with reference to flowcharts. The operation flow chart used below is CP
The U16 executes various programs stored in the ROM 18 to control the apparatus.

FIG. 3 is an operation flowchart of the Japanese-language analyzing apparatus according to the first embodiment of the present invention, showing the operation from input character string conversion to Kana-Kanji conversion and Japanese-language analysis. .

In the present embodiment, a case is considered in which Japanese characters are analyzed by performing a kana-kanji conversion on a kana character "Kare wa Amaama".

As shown in FIG. 3, when the operator instructs to start the Japanese language analyzer from the input unit 1, the control unit 13 starts the kana-kanji conversion unit 5 and the Japanese analysis dictionary unit 6. (Step 1). At this time, the control unit 13 determines whether the kana-kanji conversion device has already been activated, and if so, clears the index information for performing a dictionary search of the kana-kanji conversion dictionary unit 2 from the RAM 17. At the same time, the control unit 13 stores the kana-kanji conversion dictionary index storage unit 3 and the index information of the Japanese analysis dictionary index storage unit 7 on the analysis information storage unit 11.

Here, the index information merging unit 10
When the control unit 13 stores the kana-kanji conversion dictionary index storage unit 3 and the index information of the Japanese analysis dictionary index storage unit 7 in the analysis information storage unit 11, a character string matching the index information Is determined, and if they match, the index is merged and stored in the analysis information storage unit 11 in a form in which the matching part is removed (S2, index information merging step).
For example, in the kana-kanji conversion dictionary index storage unit 3,
For a word that is a kana character, an index into the kana-kanji conversion dictionary unit 2 in which grammatical information such as the kanji notation and part of speech of the corresponding word is described is stored (FIG. 4).

On the other hand, the Japanese analysis dictionary index storage unit 7 stores, for each word, a Japanese analysis dictionary unit 6 in which grammatical information such as the part of speech of the corresponding word and translation information necessary for translation are described. Are stored (FIG. 5). The index information merging unit 10 creates an index as shown in FIG. 6 by removing the matching part from the two pieces of index information, that is, by sharing the part of the Japanese character string serving as the index key, and storing the analysis information. It is stored in the unit 11.

Here, the kana-kanji conversion dictionary unit 2 stores the kana-kanji conversion dictionary index unit 3 as shown in FIG. In the Japanese analysis dictionary section 6, corresponding to the position information stored in the Japanese analysis dictionary index storage section 7,
It is stored like the Japanese analysis dictionary unit 6 shown in FIG.

Next, the user inputs a kana character string using the input unit 1 (S3, input step). The input character string is converted to kanji via the kana-kanji conversion unit 5 and, if it is converted for each phrase, via user intervention (S4). At this time, the kana-kanji conversion dictionary unit 2 is searched for the input character string (S5), and the kana-kanji conversion dictionary unit 2 merged in the index information merge step.
While vocabulary information that has been successfully retrieved is retrieved from the kana-kanji conversion dictionary unit 2 as needed based on the index to (S
6. Kana-kanji information acquisition step), and the kana-kanji conversion unit 5 converts the input kana character string into kanji (S7, kana-kanji conversion step).

Next, when the user gives a Japanese language analysis instruction using the input unit 1 (S8), the input character string is searched in the Japanese language analysis dictionary unit 6 (S9) and merged in the index information merging step. In addition, based on the index into the Japanese analysis dictionary, the vocabulary information that has been successfully retrieved is retrieved from the Japanese analysis dictionary unit 6 as needed (S
10, Japanese analysis information acquisition step), Japanese analysis unit 9
Performs the syntax and semantic analysis of the character string mixed with kana-kanji obtained from the kana-kanji conversion unit 5 (S11, Japanese analysis step).

Finally, the structure as shown in FIG. 7 is displayed on the display unit 12 to terminate the Japanese language analysis processing.

As described above, according to the present embodiment, a Japanese character string which is an index key for the kana-kanji conversion dictionary unit 2 and the Japanese analysis dictionary unit 6 can be shared. The memory usage during the operation of the word analyzer can be greatly reduced.

(Embodiment 2) In the first embodiment, the vocabulary information obtained from the Kana-Kanji conversion dictionary unit 2 is used in the Japanese analysis unit 9 so that the memory usage during the Japanese analysis is And the data acquisition time can be reduced. This operation will be described with reference to the flowchart of FIG.

In the second embodiment, as shown in the first embodiment, it is assumed that the processing from input of a kana character string to conversion of a kana-kanji character has been performed through S1 to S7.

Here, the kana-kanji conversion unit 5 simultaneously sends the vocabulary information used in the kana-kanji conversion to the Japanese analysis dictionary unit 6 together with the converted kana-kanji mixed character string.
At the time of Japanese dictionary search, the Japanese parsing unit 9 determines whether or not the passed vocabulary information includes information necessary for analysis from grammatical information such as part of speech in the passed vocabulary information ( S8) A Japanese analysis dictionary search is performed only for words not included (S9), and vocabulary data is extracted from the Japanese analysis dictionary unit 6 (S10). Then, the Japanese parsing unit 9 performs a syntax / semantic analysis using the vocabulary information (S1).
1) Display the result.

As described above, according to the present embodiment, in the dictionary search for Japanese analysis, the vocabulary information obtained at the time of kana-kanji conversion is used, thereby reducing the number of dictionary searches, ie, winning the Japanese analysis. It is possible to improve efficiency and prevent duplication of vocabulary information extracted from the Japanese analysis dictionary unit 6, that is, reduce the amount of memory used.

(Embodiment 3) In the first embodiment, the analysis performed by the kana-kanji conversion unit 5 at the time of kana-kanji conversion is used by the Japanese analysis unit 9 so that the Japanese-language analysis unit 9 The processing efficiency can be improved, and this operation will be described with reference to the flowchart of FIG.

In the third embodiment, as shown in the first embodiment, it is assumed that the processing from the input of the kana character string to the conversion of the kana to kanji has been performed through S1 to S7.
Here, the kana-kanji conversion unit 5 sends the syntax analysis result used in the kana-kanji conversion together with the converted kana-kanji mixed character string to the Japanese analysis unit 9 at the same time. At this time, if the user has performed kana-kanji conversion for each phrase, the result of conversion for each phrase is passed to the Japanese analysis unit 9 if one sentence has been input.

Next, the Japanese analysis unit 9 converts the kana-kanji conversion unit 5
The kana-kanji character string passed from the dictionary is subjected to a dictionary search in the Japanese analysis dictionary unit 6, and vocabulary information of the successfully searched word is extracted from the Japanese analysis dictionary unit 6. Based on the extracted vocabulary information, the analysis information passed from the kana-kanji conversion unit 5 is used in a transient state of Japanese analysis. For example, it is assumed that the analysis result passed from the kana-kanji conversion unit 5 is passed for each phrase in the following form.

Noun phrase → noun (he) + particle (wa) Noun phrase → noun (head) + particle (ga) predicate → adjective (excellent) Here, the notation A → B is the word shown on the right side This means that they form a grammatical unity phrase on the left side.

The Japanese analysis unit 9 collects the above results as phrases and as higher-level sentences, and further performs semantic analysis.

It is also assumed that the analysis result passed from the kana-kanji conversion unit 5 is passed as a sentence in the following form.

Sentence → noun phrase (he) + noun phrase (head) + predicate (excellent) The Japanese analysis unit 9 performs semantic analysis on the predicate based on the result.

As described above, according to the present embodiment, the Japanese language analysis unit 9 reuses the syntax analysis result when the kana-kanji conversion unit 5 performs the kana-kanji conversion process, thereby Since the analysis processing phase in the analysis can be skipped, an improvement in the analysis processing speed can be expected.

[0061]

As described above, according to the present invention, the input unit for inputting the kana character, instructing the start of the kana-kanji conversion and the Japanese syntax analysis, and storing the kanji corresponding to the kana character and its part of speech and syntax information are stored. An acquisition request is sent via a kana-kanji conversion dictionary unit, a kana-kanji conversion dictionary index storage unit storing the storage location of the vocabulary information of the words stored in the kana-kanji conversion dictionary unit, and a kana-kanji conversion dictionary index storage unit. The kana-kanji information acquisition unit that extracts the vocabulary information of the words from the kana-kanji conversion dictionary unit, and the kana-kanji character that converts the input kana characters into kanji codes based on the word vocabulary information obtained from the kana-kanji information acquisition unit Conversion unit, reading and pronunciation for words,
Japanese analysis dictionary that stores part-of-speech, grammar information, and semantic information, Japanese analysis dictionary index storage that stores the vocabulary information storage location of words stored in the Japanese analysis dictionary, and Japanese analysis dictionary index Based on the vocabulary information of the word requested to be acquired from the Japanese analysis dictionary unit via the storage unit and the word vocabulary information obtained from the kana kanji information acquisition unit, the kana kanji The sentence mixed with kana and kanji obtained from the conversion unit is analyzed by Japanese morphological
Japanese analysis unit that performs semantic analysis, Kana-Kanji conversion dictionary index storage unit, Index information integration unit that integrates index information in the Japanese analysis dictionary index storage unit, Kana-Kanji conversion dictionary unit index, Japanese analysis dictionary And an analysis information storage unit that stores analysis at the time of translation, kana-kanji conversion, and Japanese analysis, and a display unit that displays a Japanese analysis result. To reduce the amount of memory used when incorporating a new Japanese language analyzer into an already installed information device and using the kana information input to the kana-kanji conversion device for the semantic determination processing of Japanese language analysis. Can be planned.

Further, kana characters are inputted by the input section,
When the kana-kanji conversion unit passes the kana-kanji character string converted from kana characters to kanji to the Japanese analysis unit, the kana-kanji information acquisition unit also passes the vocabulary information acquired from the kana-kanji dictionary unit, and the Japanese analysis unit When searching the Japanese analysis dictionary, it is determined whether there is any overlapping information with the vocabulary information of the word passed from the kana-kanji conversion unit. The information acquisition unit is configured to remove the duplication from the information acquired from the Japanese analysis dictionary and acquire only the difference information of the vocabulary information. Incorporating a word analyzer and using the kana information input to the kana-kanji conversion device for the semantic determination processing of Japanese analysis, while suppressing the increase in memory usage when extracting dictionary data into memory , Ending It can reduce the data acquisition time.

Further, when the kana-kanji character is inputted by the input unit and the kana-kanji conversion unit passes the kana-kanji character string obtained by converting the kana character to kanji to the Japanese analysis unit, the kana-kanji conversion unit converts the kana character string to the kanji character. In the case of converting to Kana, the analysis result performed is also passed at the same time, and the Japanese analysis unit uses this analysis information to perform syntax and semantic analysis of the sentence mixed with Kana-Kanji output by the Kana-Kanji conversion unit. Since the word parser reuses the parsing result passed from the kana-kanji conversion and does not perform extra parsing / semantic analysis, the processing time in the Japanese parser can be reduced.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a data registration device and a translation device according to a first embodiment of the present invention.

FIG. 2 is a circuit block diagram of a data registration device and a translation device according to the embodiment;

FIG. 3 is a flowchart showing the operation of the Japanese language analyzer according to the embodiment;

FIG. 4 is a diagram showing a state of a kana-kanji conversion dictionary index storage unit in the embodiment.

FIG. 5 is a diagram showing a state of a Japanese analysis dictionary index storage unit according to the embodiment;

FIG. 6 is a diagram showing a state of an analysis information storage unit according to the embodiment;

FIG. 7 is a diagram showing a state of a kana-kanji conversion dictionary unit in the embodiment.

FIG. 8 is a diagram showing a state of a Japanese analysis dictionary unit in the embodiment.

FIG. 9 is a diagram showing the syntax and semantic structure of the Japanese analysis result

FIG. 10 is a diagram showing a syntax and a semantic structure of a Japanese analysis result.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input part 2 Kana-kanji conversion dictionary part 3 Kana-kanji conversion dictionary index storage part 4 Kana-kanji information acquisition part 5 Kana-kanji conversion part 6 Japanese analysis dictionary part 7 Japanese analysis dictionary index storage part 8 Japanese analysis information acquisition part 9 Japanese analysis unit 10 Index information merging unit 11 Analysis information storage unit 12 Display unit 13 Control unit 14 Keyboard 15 CRT 16 CPU 17 RAM 18 ROM

Claims (9)

[Claims] An input unit for inputting a kana character, a kana-kanji conversion, and an instruction to start a Japanese syntax analysis; a kana-kanji conversion dictionary unit storing a kanji corresponding to the kana character and its part of speech and syntax information; Via the kana-kanji conversion dictionary index storage unit that stores the storage location of the vocabulary information of the words stored in the kana-kanji conversion dictionary unit, and through the kana-kanji conversion dictionary index storage unit, A kana-kanji information acquisition unit to be retrieved from the kana-kanji conversion dictionary unit, and a kana-kanji conversion unit that converts an input kana character to a kanji code based on word vocabulary information obtained from the kana-kanji information acquisition unit. A Japanese analysis dictionary unit that stores reading and pronunciation for words, part of speech, grammar information, and semantic information, and a Japanese language that stores the storage locations of vocabulary information of words stored in the Japanese analysis dictionary unit A lexical dictionary index storage unit, and a Japanese analysis information acquisition unit that extracts vocabulary information of a word requested to be acquired from the Japanese analysis dictionary unit via the Japanese analysis dictionary index storage unit; and the kana kanji information. Based on the word vocabulary information obtained from the acquisition unit, a Japanese analysis unit for performing Japanese morphological analysis, syntax and semantic analysis of the kana-kanji mixed sentence obtained from the kana-kanji conversion unit, and the kana-kanji conversion dictionary index A storage unit, an index information integration unit that integrates index information in the Japanese analysis dictionary index storage unit, an index of the kana-kanji conversion dictionary unit, an index of the Japanese analysis dictionary unit, a kana-kanji conversion, and Japanese A date comprising: an analysis information storage unit for storing an analysis at the time of analysis; and a display unit for displaying a Japanese analysis result. Word analyzer. 2. A kana-kanji information acquiring unit, comprising: a kana-kanji dictionary unit for inputting kana-kanji characters through an input unit; The vocabulary information obtained from is also passed at the same time, and when the Japanese parsing unit searches the Japanese parsing dictionary part, whether there is any overlapping information between the vocabulary information of the word passed from the kana-kanji conversion part And determining if there is a duplication, the Japanese analysis information obtaining unit removes the duplication from the information obtained from the Japanese analysis dictionary unit and obtains only the difference information of the vocabulary information. Japanese analyzer described. 3. A kana-kanji conversion unit that inputs a kana character, and when the kana-kanji conversion unit passes the kana-kanji character string obtained by converting the kana character to kanji to the Japanese analysis unit, the kana-kanji conversion unit converts the kana character string into a kanji character. When converting to
2. The Japanese analysis device according to claim 1, wherein the Japanese analysis unit uses the analysis information to perform syntax and semantic analysis on the sentence mixed with the kana / kanji output by the kana / kanji conversion unit. 4. An input step for inputting a kana character, an instruction for starting kana-kanji conversion and Japanese parsing, and requesting the kana character input in the input step via a kana-kanji conversion dictionary index storage unit. The kana kanji information input step of extracting the vocabulary information of the word having the word from the kana kanji conversion dictionary unit, and the kana character input from the input step based on the word vocabulary information obtained from the kana kanji information obtaining step. The kana-kanji conversion step of converting the kana-kanji into a code, and the kana-kanji mixed character string converted by the kana-kanji conversion step, via the Japanese analysis dictionary index storage unit, the vocabulary information of the word requested to be obtained is converted to the Japanese vocabulary information. Japanese analysis information obtaining step to be taken out from the word analysis dictionary unit, vocabulary of words obtained by the Japanese analysis information obtaining step Japanese-language analysis step of syntactically and semantically analyzing a kana-kanji mixed character string converted by the kana-kanji conversion step based on the information, the kana-kanji conversion dictionary index storage unit, and the index information in the Japanese analysis dictionary index storage unit And an index information integrating step of integrating the information. 5. A kana character is inputted in an input step,
When passing the kana-kanji character string converted by the kana-kanji conversion step to the Japanese analysis step, the kana-kanji information acquisition step also passes the vocabulary information acquired from the kana-kanji dictionary part,
When the Japanese analysis step searches the Japanese analysis dictionary part, it is determined whether or not there is overlapping information with the vocabulary information of the word passed from the kana-kanji conversion step. 5. The method according to claim 4, further comprising the step of: removing duplicates from the information acquired from the Japanese analysis dictionary unit and acquiring only difference information of vocabulary information. . 6. A kana character is inputted in an input step,
When passing the kana-kanji character string converted by the kana-kanji conversion step to the Japanese analysis step, the analysis result performed when the kana-kanji conversion step converts the kana character string to kanji is also passed at the same time. Using the analysis information,
5. The Japanese analysis method according to claim 4, further comprising: a Japanese analysis step of syntactically and semantically analyzing the sentence mixed with the kana / kanji output by the kana / kanji conversion step. 7. An input step for inputting a kana character, instructing the start of kana-kanji conversion and Japanese syntax analysis, and requesting the kana character input in the input step via a kana-kanji conversion dictionary index storage unit. The kana kanji information input step of extracting the vocabulary information of the word having the word from the kana kanji conversion dictionary unit, and the kana character input from the input step based on the word vocabulary information obtained from the kana kanji information obtaining step. The kana-kanji conversion step of converting the kana-kanji into a code, and the kana-kanji mixed character string converted by the kana-kanji conversion step, via the Japanese analysis dictionary index storage unit, the vocabulary information of the word requested to be obtained is converted to the Japanese vocabulary information. Japanese analysis information obtaining step to be taken out from the word analysis dictionary unit, vocabulary of words obtained by the Japanese analysis information obtaining step Japanese-language analysis step of syntactically and semantically analyzing a kana-kanji mixed character string converted by the kana-kanji conversion step based on the information, the kana-kanji conversion dictionary index storage unit, and the index information in the Japanese analysis dictionary index storage unit And an index information integrating step of integrating the information. 8. A kana character is input in an input step,
When passing the kana-kanji character string converted by the kana-kanji conversion step to the Japanese analysis step, the kana-kanji information acquisition step also passes the vocabulary information acquired from the kana-kanji dictionary part,
When the Japanese analysis step searches the Japanese analysis dictionary part, it is determined whether or not there is overlapping information with the vocabulary information of the word passed from the kana-kanji conversion step. 5. The Japanese analysis program according to claim 4, further comprising: a Japanese analysis information obtaining step of obtaining only difference information of vocabulary information by removing duplication from information obtained from the Japanese analysis dictionary unit. Recording medium on which is recorded. 9. A kana character is input in an input step,
When passing the kana-kanji character string converted by the kana-kanji conversion step to the Japanese analysis step, the analysis result performed when the kana-kanji conversion step converts the kana character string to kanji is also passed at the same time. Using the analysis information,
5. The recording medium according to claim 4, further comprising a Japanese language analyzing step for syntactically and semantically analyzing the sentence mixed with the kana / kanji output by the kana / kanji converting step.