JPH0869463A - Device and method for japanese syllabary and chinese character conversion - Google Patents

Abstract

PURPOSE: To obtain a desired paragraph space-interposed writing candidate by making modifications in consideration of causative passive cases and derivative notation and precisely dividing paragraphs with interposed spaces. CONSTITUTION: For a paragraph space-interposed writing process, the words of the latter paragraph are noticed and it is judged whether or not the word has modification information (step S300); when an attached word is not allowed although the corresponding modifying word is present, a causative and passive modification verifying process is performed (step S342). For causative and passive cases, allowed attached words change, so this is verified and when the attached word is allowed, the modification is formed and this is regarded as a paragraph candidate; and the range from the modified word to the modifying word is excluded from the retrieval range on and after next-time modifications (steps S350 and S360). Consequently, the formation of modifications is decided in consideration of the causative and passive cases and the possibility that a desired paragraph space-interposed writing is obtained can be improved. Further, the modifications are verified by using only representative notation.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a kana-kanji conversion device and kana-kanji conversion method. Kana-Kanji conversion method.

[0002]

2. Description of the Related Art Conventionally, various kana-kanji conversion devices for converting a kana character string input from a keyboard into a desired kana-kanji mixed sentence have been proposed as an input device for Japanese sentences or an editing device for Japanese sentences. Has been done. As such a kana-kanji conversion device, it is desired that the user does not need to specify the delimiter positions of words and phrases one by one, and that the converted character string has the notation desired by the user. In Japanese, since there are many homophones and homonyms, in order to obtain the desired mixed kana-kanji sentence without error, the meaning of the sentence must be finally analyzed. Analysis requires a knowledge base of at least tens of thousands of organically related words,
Realization is extremely difficult.

Therefore, in the conventional kana-to-kanji conversion device, devising the process of segmenting and writing bunsetsu and the process of selecting homonyms and oppositions, an attempt is made to obtain the desired result without analyzing the meaning. . The bunsetsu segmentation processing includes a 2-bunsetsu longest matching method in which a bunsetsu that gives the longest bunsetsu that can be formed with two bunsetsu as a basic unit is the first candidate, and a word that can be a candidate for a word forming a bunsetsu There is a minimum cost method in which a cost is added to a combination of words and a phrase whose score satisfies a predetermined condition is set as a first candidate. In the learning process, the most recently selected homophonic word from the homonym is used as the next candidate for learning the homonym object, and the length of the phrase containing a word is used immediately before. It is known to learn bunsetsu length, which gives top priority to the length specified by the person.

Furthermore, recently, attention has been paid to a specific relationship between words (for example, "hot" and "tea" in "hot tea" or "hot" and "summer" in "hot summer", and this relationship is considered. By preparing a dictionary that stores, when one word (eg, “tea”) is specified, a word related to this word (eg, “hot”) is selected as the first candidate. (For example, “Kana-Kanji conversion device” of Japanese Patent Laid-Open No. 3-105664 and “Kana-Kanji conversion device” of Japanese Patent Laid-Open No. 4-2777861). The particular relationship between these words is called "dependency" or "co-occurrence."

[0005]

However, these kana-to-kanji conversion devices only look at the relationship between words in a bunsetsu obtained by performing bunsetsu segmentation, so if the segmentation syllabary is incorrect. , A dictionary that remembers the relations between the prepared words is also useless.
In the first place, since we were only looking at the relationship between adjacent words as the relationship between words at most, there were cases where we could not obtain the most natural kana-kanji mixed sentence in Japanese. However, if the scope of studying the relationship between words is unnecessarily widened, the number of combinations will increase in a geometric progression as the number of input kana characters increases. By the time it takes an unacceptably long time.

Further, regarding such a relationship between words, in the information that two or more words are simply used in close proximity (hereinafter referred to as "co-occurrence" in a broad sense), the role of a particle in Japanese is It is overlooked and the correct conversion result cannot be obtained. For example, even if “Kishakushasha” can be correctly converted to “Reporter returns to work” by simply storing the information that “reporter” and “return to work” have a co-occurrence relationship, "I will return to the reporter"
Will not be the correct conversion. Therefore, as the information on the modification, the relationship between the modification word and the modification word, such as "noun" + "particle" + "synonym" (for example, "reporter" + "ga" or "ha" + "return to work") It may be possible to memorize this as dependent information that includes the information of adjuncts (particles, etc.) that are allowed in, but in the case of Japanese, in the expressions such as passive and causative, the adjunct particles are changed. In order to do this,
Invites a problem that cannot properly convert "Crush". However, if the range of allowable adjunct words is expanded in order to recognize the establishment of dependency in the case of passive or causative, the accuracy of kana-kanji conversion by dependency will decrease.

Further, in the case of Japanese, there is a feature that different notations are allowed even for the same word, such as "acceptance" and "acceptance", or "acceptance". There was a problem that if all the changes were to be accepted, not only all the words in the notation should be registered in the word dictionary, but also in the change dictionary. If all the derivative notations are registered in order to make all these derivative notations subject to dependency, not only will the capacity of the dependency dictionary become extremely large, but it will It was also thought that the time required would increase and would be unacceptable.

The kana-kanji conversion device and the kana-kanji conversion method of the present invention solve these problems and provide a dependency relationship for separating the input character string and changing the priority of kanji candidates in consideration of causative and passive. It was done for the purpose of obtaining a desired mixed Kana-Kanji sentence and adopted the following structure.

[0009]

A kana-kanji conversion device of the present invention inputs a kana character string, refers to a word dictionary, and scribes the inputted kana character string into bunsetsu segmentation candidates. A kana-kanji conversion device that creates a candidate character string that forms a kana-kanji mixed sentence using the bunsetsu segmentation candidate, and provides information on a dependent word and a dependent word that form a modification between predetermined phrases. The dependency information dictionary stored together with the information of the permissible ancillary words permitted between the dependency word and the dependent word, and when performing the process of segmenting the input character string into phrases, refer to the dependency information and A phrase search means for searching a phrase including a word corresponding to the information of the dependent word and the dependent word, and attached to the related word when the phrase including a word corresponding to the information of the related word and the dependent word is searched The word to do is before First determines whether the allowable comes word 1
When the phrase including the word corresponding to the information of the related word and the related word is searched, and the attached word following the received word is a word indicating causative or passive, the related word Second determining means for determining whether a word attached to the word is a word corresponding to causative or passive, and a clause for limiting the phrase segmentation candidate based on the determination results of the first and second determining means. The gist is that the candidate limiting means is provided.

A kana-kanji conversion method corresponding to this kana-kanji conversion device inputs a kana character string, refers to a dictionary, and scribes the kana character string input, and generates kana-kanji mixed character string candidates. Kana-Kanji conversion method
When performing the process of segmenting the input character string into bunsetsus, refer to the dependency information in which the information of the dependency between predetermined phrases is stored together with the information of the admissible adjunct word permitted between the dependency word and the received word. Then, a phrase having a word corresponding to the dependency information is searched, and a phrase having a word corresponding to the dependency word and the information of the dependency word is searched and found,
It is determined whether or not the word attached to the dependent word is the allowable attached word, and when the attached word following the received word is a word representing causative or passive, the word attached to the dependent word is a causative or The gist is to determine whether or not the word is passive and to limit the candidates for the phrase segmentation based on the determination result of any one of them.

According to the first kana-kanji conversion device and kana-kanji conversion method of the present invention configured as described above, a dependency word and an adjective which constitute dependency of a predetermined phrase in the dependency information dictionary are provided. The information is stored together with the information of the permissible adjunct words allowed between both words, and when performing the processing of writing the input character string into phrase segments, refer to the dependency information stored in this dependency information dictionary. , Search for a phrase having a word corresponding to the dependency information. When a phrase corresponding to the information of the dependent word and the dependent word is searched, it is determined whether the word attached to the dependent word is an admissible adjunct word, and the adjunct word following the dependent word is a word indicating causative or passive. Has
It is determined whether the word attached to the related word is a word corresponding to causative or passive. Based on this search result, candidates for phrase segmentation are restricted. Therefore, when the dependency information exists, the dependency is not established simply by the existence of both words, but the dependency is established when the word attached to the dependency allows the relationship between the two terms, In addition, if the received word is a word indicating causative or passive, and if the word attached to the dependent word is a word corresponding to causative or passive, then the candidate candidates are restricted as the establishment of the dependent word. Undesired candidates for segmentation are less likely to be selected, and the possibility of desired segmentation is increased. If the priority of the kanji candidates for each bunsetsu is changed instead of the restriction of the bunsetsu candidates based on the bunsetsu segmentation already estimated by another method, the desired kanji candidate is set as the first candidate. The possibility of being obtained can be increased.

Here, when the dictionary is a word dictionary storing a reading of a word and a notation corresponding to the reading, and when there are a plurality of notations corresponding to the reading of the word, a representative A notation defined as a notation and a notation information unit that stores a notation determined as a derivative notation are provided, and the phrase search means searches the dependency using only the representative notation stored in the word dictionary. And means for displaying the converted candidate character string using the representative notation and the derivative notation stored in the notation information section of the word dictionary for the phrase segmentation candidate. Can be

In this kana-kanji conversion device, the representative notation and the derivative notation are stored in the word dictionary, and regarding the dependency test for each phrase segmented in the first kana-kanji conversion device, The word is specified by using only the representative notation, and the candidate character string for the specified word is displayed using the representative notation and the derivative notation.

Further, in the above kana-kanji conversion device, instead of the phrase candidate limiting means, the kanji characters for changing the priority of the kanji candidates for each phrase based on the judgment results of the first and second judging means. Candidate priority means may be provided. In this case, those that do not correspond to the dependency are left as the Kanji candidates, but those that are dependent on the Kanji are prioritized, which is preferable.

Further, the phrase searching means in the kana-kanji conversion device uses the backward phrase as a starting point and excludes the already-searched range that has already been registered, and sequentially obtains phrases including words corresponding to the dependency information toward the front. Retroactive search means to search,
When a phrase having a word corresponding to the dependency information is found by the search, an already-searched range registration means for registering the found phrase from the starting phrase to the found range of the dependency information It can also be equipped with. In this way, at the time of the next search, this range is excluded from the search range, so that the search time for separating words can be shortened, and further, the chained dependency will not be erroneously selected.

When the first or second judging means judges the existence of the relevant related word, the received word and its adjunct word, the phrase candidate limiting means in the kana-kanji conversion device described above is used to write a phrase segment containing the word. It is preferable to provide a means for preferentially selecting, because priority is given to the segmentation for which dependency is established.

In the kana-kanji conversion device equipped with the kanji candidate priority means, the kanji candidate priority means is used by the first or second determination means to determine the existence of the relevant related word, the received word and its adjunct word. In this case, it is possible to provide means for preferentially selecting a phrase segmentation including the word and selecting the word as a first candidate for kana-kanji conversion. In this case, the segmentation and the selection of the first candidate are made based on the dependency, and the establishment of the dependency has the highest priority.

Further, when the auxiliary word existing between the words having the dependency relation is a word having a predetermined specific grammatical structure, the phrase searching means determines that the dependency relation is established. By providing a means for doing so, it becomes possible to reduce the amount of information about the attached word.

Further, in the kana-kanji conversion device equipped with the phrase search means, the phrase search means refers to the dependency information stored in the dependency dictionary starting from a predetermined phrase and falls under the dependency information. When a bunsetsu equipped with a word that is found by the bunsetsu bunsetsu search unit and a bunsetsu bunsetsu search unit that finds a bunsetsu having a word A range from a phrase to the found phrase may be provided with a search range excluding unit that excludes the range from the next alternate phrase searching unit.

In this case, it is possible to eliminate the crossing of the dependencies and avoid the erroneous determination of the correct dependencies and to speed up the search for the dependencies.

In such a kana-kanji conversion device, a known phrase segmentation method can be applied to a range in which a phrase including a word having a dependency relation is not found by referring to the dependency dictionary. . For example, the two-phrase longest matching method may be used, or a score (cost) is attached to the easiness of the word-to-word or / and bunsetsu-joining so that the easiness of this joining becomes maximum (minimum cost). Words and / or phrases may be selected. It is desirable to select a combination that maximizes the likelihood of word-to-word connection and bunsetsu-to-word connection.

A second kana-kanji conversion device of the present invention inputs a kana character string, writes out the input kana character string for each phrase, and forms a kana-kanji mixed sentence corresponding to the input kana character string. A kana-kanji conversion device for generating a candidate character string, which is a word dictionary storing a reading of a word and a notation corresponding to the reading, and there are a plurality of notations as the notation corresponding to the reading of the word. In this case, a word dictionary having a notation information section that stores the notation determined as the representative notation and the notation determined as the derivative notation, and the information of the dependent word and the received word that form the modification of predetermined phrases , A stored dependency information dictionary, a segmentation processing unit for performing segment segmentation processing using the word dictionary, and for each segment segmented, only the representative notation stored in the word dictionary is used. And a grammar processing unit that performs a dependency test using the dependency information dictionary to identify a word that constitutes the input character string, and a notation information unit of the word dictionary for the identified word. The gist of the present invention is to provide candidate character string display means for displaying the candidate character string after conversion, using the representative notation and the derived notation stored in.

The invention of a kana-kanji conversion method corresponding to such a kana-kanji conversion device is to input a kana character string, write the input kana character string into punctuations, and at the same time, enter a kana kanji character corresponding to the input kana character string. A kana-kanji conversion method for generating a candidate character string that forms a mixed sentence, wherein a plurality of notations are provided as a notation corresponding to the reading of a word in a word dictionary that stores the reading of the word and the notation corresponding to the reading. If it exists, the notation defined as the representative notation and the notation defined as the derivative notation are stored, and the dependency information dictionary stores the dependency word and the information of the dependency word that constitute the dependency of the predetermined clauses. Storing and performing phrase segmentation processing using the word dictionary, using only the representative notation stored in the word dictionary for each segment segmented, using the dependency information dictionary A dependency test is performed to identify the words that make up the input character string, and using the representative notation and the derivative notations stored in the word dictionary for the identified words, the converted candidate characters The main idea is to display columns.

In this kana-kanji conversion device and kana-kanji conversion method, when the dependency check is performed by referring to the dependency information dictionary, only the representative notation registered in the word dictionary is used for the validation and the input is performed. The words constituting the character string are specified. When displaying the candidate character strings of the specified words, not only the representative notation stored in the word dictionary but also the derivative notation is used for display. Therefore, it is possible to achieve both high-speed grammar processing including dependency and variety of notation.

[0025]

Preferred embodiments of the present invention will be described below in order to further clarify the structure and operation of the present invention described above. FIG. 1 is a block diagram showing a control logic for kana-kanji conversion, and FIG. 2 is a block diagram showing hardware on which the kana-kanji conversion control logic actually operates. As shown in FIG. 2, this device is provided with the following parts centered around a well-known CPU 21 and mutually connected by a bus 31.
Each part mutually connected by the CPU 21 and the bus 31 will be briefly described.

ROM 22: Mask memory for storing kana-kanji conversion program, RAM 23: Readable and writable memory constituting main memory, Keyboard interface 25: Interface for controlling key input from keyboard 24, CRTC 27: Color display A CRT controller that controls the output of signals to the CRT 26 that is possible, a printer interface 29: an interface that controls the output of data to the printer 28, a hard disk controller (HDC) 30; an interface that controls the hard disk 32. The hard disk 32 stores various programs loaded and executed in the RAM 23, a kana-kanji conversion processing program provided in the form of a device driver, various conversion dictionaries referred to by the kana-kanji conversion processing program, and the like.

With the hardware thus configured,
Text is input, Kana-Kanji conversion, editing, display, printing, etc. That is, the character string input from the keyboard 24 is subjected to predetermined processing by the CPU 21, and the RAM 2
3 is stored in a predetermined area, and a CRT is sent via the CRTC 27.
It is displayed on the screen of 26.

Next, the function executed by the hardware thus configured will be described with reference to FIG. The configuration and operation of each unit shown in FIG. 1 will be briefly described. The processing performed here is executed by the central processing unit (CPU 21) based on the data input from the keyboard 24. The CPU 21 performs all processing. Regarding kana-kanji conversion, when the keyboard 24 is operated, a predetermined interrupt process is activated, the input key image is converted into a corresponding kana-character string, and further this is converted into a kana-kanji mixed character string. Will start. Of course, if the computer can process in parallel,
Kana-Kanji conversion may be performed by one application (input method), and the conversion result may be delivered to the required application.
In this case, the input method collectively accepts the input from the keyboard 24.

The key image from the keyboard 24 is accepted by the character input unit 40, and converted into a corresponding kana character string here. When inputting romaji, it refers to a predetermined conversion table and converts into a kana character string. Every time one kana character is obtained, the character input unit 40 sends the kana character to the conversion control unit 42. This conversion control unit 42
Is a central part of kana-kanji conversion, controls various kana-kanji conversions described later, and sends the result to the converted character string output unit 44. Converted character string output unit 44
Actually sends a signal to the CRTC 27,
Display the converted string.

The conversion control unit 42 delivers the received kana characters to the character string input unit 50. The character string input unit 50 stores the kana character string in the character storage unit 52. Based on this character string, the independent word candidate creation unit 54 and the adjunct word candidate creation unit 64
And create word data candidates. The independent word candidate creation unit 54 uses the independent word dictionary 58 stored in advance in the hard disk 32, and under the management of the independent word analysis position management unit 56, performs processing for extracting independent word candidates from the obtained kana character string. To do. On the other hand, the adjunct word candidate creating unit 64 also performs the process of extracting an adjunct word candidate from the obtained kana character string under the control of the adjunct word analysis position managing unit 66 using the adjunct word dictionary 68. The process of extracting independent word candidates and adjunct word candidates while moving the analysis position will be described later.

Here, the independent word dictionary 58 is learned by learning.
Change the priority of homonyms and affixes. The learning process is performed by the dependency learning unit 70, the independent word learning unit 72, the auxiliary word learning unit 74, the affix learning unit 76, and the character conversion learning unit 78. When the user selects a word other than the word corresponding to the dependency under the condition that the dependency is satisfied, the dependency learning unit 70 gives priority to the combination selected by the user in the combination of the same words. This is to learn the relationship of dependency. The independent word learning unit 72 learns the word selected last in the independent word group in which homonyms exist as the highest priority candidate. The auxiliary word learning unit 74 learns in which word form an auxiliary word such as "please" is converted to "please" or "please". Further, the affix learning unit 76 is for learning conversion formats such as prefixes and suffixes (for example, "Go" and "Go"). The character conversion learning unit 78 learns the input character string as it is as hiragana or katakana and outputs the confirmed hiragana or katakana as a candidate in the subsequent conversion process. .

The independent word candidate creating unit 54 and the auxiliary word candidate creating unit 64 obtain the created word candidates, and the word data creating unit 80 creates data for each word candidate. That is, as a result of the connection verification unit 82 referring to the connection verification table 84, the connection between the obtained independent words and adjunct words, independent words and independent words, and further the connection between the clauses consisting of “independent words + adjunct words”,
And the cost calculation unit 86 obtains the result of the cost calculation performed by the cost calculation unit 86 and outputs the result as data for each word.
This word data is once stored in the word data storage unit 100, receives the adjustment output from the dependency candidate adjustment unit 90, and is used for the phrase segmentation writing process.

The dependency candidate adjusting section 90 receives the word candidates from the independent word candidate creating section 54 and the adjunct word candidate creating section 64 and tests the dependency. The modification check is performed by referring to the modification dictionary 98 prepared in advance on the hard disk 32. The range in which the dependency check is performed is managed by the dependency range management unit 96. In addition, there are some admissible conditions in the verification of the dependency relationship, and these are determined by the causative / passive analysis unit 92, the particle admission analysis unit 94, and the like. The dependency candidates adjusted by the above dependency test and the word data described above are integrated by the word data storage unit 100 and provided to the phrase segmentation and segmentation processing by the segment segmentation and segmentation unit 102. The segment segmentation writing unit 102 determines a first segment segmentation segmentation candidate from the obtained data.

By the above processing, the first candidate for segmentation and segmentation and the first candidate for kana-kanji conversion for each segment are determined. The phrase segmentation writing unit 102 stores the candidates in the phrase data storage unit 106, and the stored candidates are output to the conversion control unit 42 by the conversion character string output unit 108. The conversion control unit 42 displays this character string as a candidate character string, and since an undesired character string may be a candidate, the conversion control unit 42 receives instructions from the user and performs processing such as display and selection of the next candidate. To do. The results of these instructions and selections are
It is input to the phrase data storage unit 106 and each of the learning units 70 to 78 described above, and is used for partially fixing the phrase and rewriting the priority order by learning. Although not shown, when the character string is confirmed by the user, all the data temporarily stored in each unit are erased and prepared for the next conversion.

The process from the input of kana characters to the output of the conversion word character string has been outlined above. Next, the details of each process will be described. First, a general phrase segmentation process will be described, and then a dependency process, which is an essential part of the present invention, will be described. FIG. 3 is a flow chart showing an outline of the processing of phrase segmentation by the minimum cost method. As shown in the figure, first, initialization processing such as erasing of temporarily stored data and initialization of the analysis position in the first digit (step S200) is performed, and then processing for obtaining the analysis position is performed ( Step S210). The analysis position is a position where the kana character string input so far is advanced one by one from the beginning. For example, if a kana character string “Kuru ma wa ko wo ko bu ko” is input as in the example sentence shown in FIG. 4, the first analysis position is the position of “KU” in the first digit. At this analysis position, a process of searching the independent word dictionary 58 and the auxiliary word dictionary 68 stored in the hard disk 32 is performed (step S220).

After the dictionary is searched, the obtained word is subjected to a process of checking the combination with the previous word (step S230). Further search the dictionary.
For example, in the example shown in FIG. 4, "ha" of "koohakobu"
The particle "ha" retrieved from the adjunct dictionary 68 for
Is determined to be impossible to combine with the immediately preceding case particle “o”, and is treated as invalid data by the connection test by the word data creation unit 80 and the connection verification unit 82. In FIG. 4, a word "x" is added to a word judged to be invalid by such a combination check. The connection between words is stored in advance in the connection verification table 84. This connection verification table 84 is a table that gives information about the possibility of combining the parts of speech of words. In the embodiment, 400 is used. It is given as a matrix of about × 400. When the dictionary search and the combination check at one analysis position are completed, the analysis positions are sequentially advanced and the processing is repeated.

For words that may be combined, the cost is calculated next, and the minimum total cost of the words is calculated (step S240). This processing is performed by the cost calculation unit 86, and in the example shown in FIG.
"Car" can be divided into, for example, "ku" + "ru" + "ma", "car" + "ma" and "car". When applying words to these, independence word = 2, attached word It is assumed that the cost is = 0, and if "bitterness" (independent word) + "stream" (independent language), the total cost of "stream" is 4. At this time, the cost of "pause" is 4 because the minimum total cost is obtained, and "pain" + "flow" + "pause"
This is because the cost 4 in the case of “come” + “between” is adopted instead of the cost 6 of. Since "de" and "ha" are subordinate words, the word "car" with the lowest cost among the words before it =
The cost of 2 is its own cost. In FIG. 4, the cost of each word is shown in the lower right.

After the cost calculation described above, the cost of each word is checked, and the process of invalidating the cost is performed (step S250). What is an inappropriate cost?
It is a combination of words that costs more than other combinations of words. That is, selecting a combination of words such as "ku" + "dome" is more expensive than other words "come" and "repeat" that can be obtained up to that position. Judgment is made and this is excluded from the phrase candidates. Words that are not adopted due to the idea of this minimum cost are shown as “●” in the upper right of the words in FIG. Note that, in FIG. 4, “◯” indicates that the word remains as a word that may form a phrase candidate as a result of the above-described combination check and cost check.

Next, a process of linking the word candidates thus given the cost is performed (step S260). That is, with respect to the words for which the combination is valid, the connection relation is related by setting a pointer. In the example of FIG. 4, "come", "roll", "car", "up", "de", "ha"
Since the minimum total cost was calculated as a word such as "in" which is not invalid, "up" is "" for "come" and "iteru".
The relationship is such that "car" is linked to "de" and "de" is linked to "car". The processes of the combination check, the cost calculation, and the linking are repeated while all the words are searched at one analysis position. Also, when the dictionary search at the analysis position is completed,
Further advance the analysis position, examine the establishment of a new word, and repeat the join check and cost calculation in the same way.

When the analysis position has reached the position of the last kana character that has already been input and the analysis has been completed for all words (step S265), the process of searching for the path with the minimum cost based on the above process. (Step S27
0). This processing is performed by the phrase segmentation writing unit 102, and is a processing of searching for a combination of valid words that minimizes the total cost given to the words. In the case of "Walking until you come", as shown by the path of the solid line J in Fig. 4 (B), "By car" + "Box" +
The total cost is 18 for the "carry" segment.
Selected as the lowest cost. It should be noted that although the cost is not the minimum, other phrase segmentation candidates are also searched. For example, as shown by a path indicated by a broken line B in FIG. 4B, the segmentation is “by car” + “child” + “carry” (cost = 20). After creating the segmentation candidate in this way (step S280), this time, the process of creating the candidate inside each phrase is performed (step S290). That is,
Within a bunsetsu syllabary, candidates such as "box" and "box" for "hako" are prepared. These bunsetsu candidates and word candidates are used when the user instructs to change the way of bunsetsu division or to display the next candidate.

FIG. 5 shows an example of another phrase segmentation. Although this example is used for the description of the dependency described later, the phrase segmentation by the minimum cost method that does not take the dependency into consideration is applied to “make a choke”. In this example, the combination check (step S230) excludes “ya” from the fact that the direct combination with the noun is invalid (marked with “x”). Further, “tree”, “Mr.”, “company”, etc. are excluded from the check of the minimum total cost (step S250) (marked with ●). As a result, in the example shown in FIG. 5, "kisha o" + "kisha"
Phrase segmentation is done and the priority of each word is shown in Figure 5.
If the order is as shown in (1), "your company" will be selected as the first candidate for "kisha." As for the latter half of “Kisha”, if the analysis reaches only the end of “Kisha”, the same “Your company” as the first candidate in the first half will be selected. When "sease" or "suru" is attached after "," for example, "return to work" considering that the word "sease" is attached
Can be displayed as the first candidate.

In the above description, the cost calculation is performed only for each word itself, but in reality, points are added to reduce the cost according to the degree of easiness of combining words, and clauses are added. It is also possible to add points for reducing the cost to the combination of clauses which are also easy to combine based on the grammatical rules regarding the connection between them. Here, for the convenience of understanding the processing of phrase segmentation,
It is explained using the simplest method.

Based on the above-described processing of segmentation and segmentation, the processing of segmentation and segmentation by dependency will be described below. FIG. 6 is a flow chart showing the process of performing the dependency check. This processing is performed in parallel with the processing of steps S220 to S250 shown in FIG. Actually, after obtaining the analysis position (step S210), when searching various dictionaries, the independent word dictionary 5
8. Dependent dictionary 98 along with the search of the attached word dictionary 68
Then, the following dependency check process is performed together with the join check and the cost check accompanying the calculation of the minimum total cost. When this process is started, first, the dependency candidate adjusting unit 90 searches the dependency dictionary 98 for words that can be candidates at the analysis position (words marked with a circle) to determine whether the dependency information exists. Determine whether or not (step S
300). If the word at the analysis position is a conjunction, a verb, or an independent word, there is no dependency,
The process for that word ends immediately.

As an example, a case will be described in which the kana character "Kisosha Kisesha" is input and the analysis proceeds to "Kisasha Kisha". At this time, as a candidate for "Kissha" in the latter half, "reporter""yourcompany""train"
Since you can get "return to work" etc., for each of these words,
It is checked whether there is any information in the dependency dictionary 98. In addition, when the word to be received is a verb such as "heard" or "taku" and has an inflectional form, use the stems "hear" and "li" or the basic forms "hear" and "difficult" as keywords. Thus, the dependency dictionary 98 is configured so that it can be referred to.

FIG. 7 shows an example of the structure of the dependency dictionary 98 in the embodiment. The dependency dictionary 98 of this embodiment is
The structure has a unit of [heading + word (stem) + 1 related word + adjunct information]. In the example of FIG. 7, for the word "return", the heading "kisha" + word "return" ”+ Word“ reporter ”+“ ga ”, headline“ kisha ”+ word“ return to work ”+ word“ yousha ”+“ ni ”, headline“ kisha ”+ word“ return ”+ word“ train ” "+" De "...
・ ・ There are multiple sets of data for one word. Further, for the heading “Kisha” and the received word “reporter”, for each of the words that compose the related word group “Your company, newspaper, news agency ...” It has data for each related word, such as "+ related word" Your company "+" no ". These data are arranged in the order of the Japanese syllabary of the headword for the word. Of course, similar dependency information is stored for other words. Dependency candidate adjustment unit 90
Searches this dependency dictionary, and if the corresponding heading and the relevant word exist, takes out the candidate of the dependency word from the dictionary and uses it for the dependency verification. Since these data have the same headline and the same word, the headword and the word may be prepared at the head of the entire data group, and the related word and the related word information may be prepared individually.

Further, in the example of FIG. 7, only the minimum information is shown in order to clearly show the dependency check. "Word heading length", "received word kanji",
It includes detailed information such as "accept word part of speech", "engagement word head length", "engagement word kanji", "engagement word part of speech", "engagement relationship". The part-of-speech of the dependent word and the dependent word is necessary for considering the establishment of the dependent word and the admissibility of the dependent word, and the heading length data is necessary for the high-speed extraction from the dictionary 98.

When a word for which dependency information is present (hereinafter referred to as a "bearing word") is found, next, except for the range in which the dependency is already established, the dependency is moved forward. A search is made for a corresponding word (hereinafter referred to as a dependent word) (step S310), and it is determined whether or not there is a corresponding related word (step S32).
0). At this time, the related word is searched for not only the word having the minimum total cost but also other words. Now, in the dependency dictionary 98, as shown in FIG. 7, the relations of “reporter (return)”, “your company (return)”, “your reporter”, and “train (re) return” It is assumed that the receiver is stored. Here, the kana in () is an adjunct word that is permitted as possibly existing between words that have a dependency relationship. For "return to office" and "reporter" that are determined to be possible words in step S300, there is a word in the dependency information. Therefore, when it is determined whether or not there is a dependency word in each of these, in the example shown in FIG. For “return to work”, it is determined that the candidates for the character string “kisha” are “reporter”, “your company”, and “train”, and for “reporter”, the candidate for the character string “kisha” is “your company”. Is determined to be applicable. If it is determined in step S320 that the dependent word exists,
Next, it is judged whether or not the adjunct word existing between the two words is a word that allows the existence of the dependency (step S33).
0).

The adjective parsing of particles determines whether or not the admissibility relation defined by the type of dependency is observed, and has the following types for each type of dependency. [I] Consecutive modifier noun + particle + particle in case of adjective Case particle “ga” “kara” “de” “to” “ni” “e” “yori” “o” “no” particle “ha” In the case of the noun + noun + noun + noun + noun (particle abbreviation type), the optional particles "ga", "ha", and adverbs [II] adnominal modifier noun + particle + noun particle In the case of "no" body language + body language (parallel) "ya""to" in the case of conjunctions + nouns In the case of conjunctions + nouns

That is, it is determined that the dependency relationship is 2
If the relation between two words belongs to any of the above or any of the above, and if the adjunct word (mostly a particle or particle-like expression) that exists between the two words in the dependency relationship falls under any of the above, The dependency dictionary 98 is set with a particle that allows a word having a dependency relationship, and this is verified. For example, the dependency between "tact" and "taku" is the setting of the particle allowance (no ・ ga)
If it is accompanied by, it belongs to the above case (noun + particle + noun), so "no" and "ga" can exist between both words (quickly, witty → ○) , Other case particles such as "kara" and "de" are unacceptable.

Regarding each of the relationships (1) to (4), it is determined that the relationships other than those listed therein are allowed. The examples of judging that this is permissible are listed below, but these may include cases in which the dependency does not hold as an actual expression. However, dependency is a broad concept as an actual human language activity, and too strict dependency arrangements are often not practical. Further, too strict dependency arrangement only causes an unnecessarily large increase in the dependency dictionary 98, and also reduces the speed of dependency verification. Therefore, in this embodiment,
Regarding the admissibility of ancillary words, the relationship in which the dependency is generated is separated from that, and if there is a clear allowance / non-allowance, it is stored in the dependency dictionary along with the words that have the dependency relationship, and otherwise. With regard to

[III] Permissible expression-in the case of continuous modified form ・ Noun + case particle expression + case particle expression in a noun "each""as""for""in""by", etc. Noun + particle + particle particle in adjectives “koso,” “sae,” “shika,” “but,” “mo,” etc. ・ Noun + auxiliary particle + auxiliary particle in adjectives “cut”, “about”, “gatsu”, “only” Etc. ・ Noun + Adverbial expression + Adverbial expression in adjectives "because", "Nara", etc. ・ Adjective + Particles + Particles in adjective "Noha", etc. "Kara", "te", etc. ・ Connective particle-like expressions "kara ni""wa" and "ho"
"Ue", etc. ・ Expression in which verbs and verbs are arranged in parallel "ka""shi""tari"
"At the same time", etc.

[IV] Permissible expressions-adjective modifiers-Noun + particle-like expressions + particle-like expressions in nouns "in", "relating", "based on", etc. Expressions such as “for”, “to”, “to accompany”, “to”, “toki”, etc.

In accordance with the above rules, the admissibility of the adjunct word between the two words for which the dependency relationship is found is determined. In the case of "reporter" and "return to work" given as an example, the allowable case particle is "ga", so "kisha ki kisha"
Regarding the above, the acceptance of the dependency cannot be recognized. Therefore, this is determined (step S340), and if it is determined that the dependency is not established despite the presence of the dependent word and the dependent word that establish the dependency, then the causative / passive relationship is determined. Acceptance verification processing is performed (step S34).
2).

The causative / passive dependency verification process is performed by the causative / passive analyzing unit 92 shown in FIG. This process will be described in detail. In the example sentence shown in FIG. 5, when the analysis further progresses to “Kissakusha” as shown in FIG. 8, it can be determined that it is a causative / passive, and the dependency processing considering the causative / passive case is performed. It will be performed (step S342). This process is performed in steps S320 and S330.
Similarly, there is a point that there is a corresponding related word and a judgment that an adjunct in that case is an allowable word.
Focusing on “returning” to “kisha”, the corresponding word “reporter” exists, and when the adjunct is analyzed next, in the case of causative, the original adjunct “ga” is “ Since it is stored in advance that "is allowed", it is determined that the dependency is established. Note that the relationship between "return to work" and "your company" is that the originally admissible adjunct is "ni", and "wo" is not allowed even in the case of causative, so the relationship is established. It is not allowed. Similarly, "return to office" and "train" + "de" are also tested for dependency, and it can be seen that there is no adjunct that is allowed as a causative between the term and the word.

Therefore, it is judged based on these analysis results whether or not the dependency is established (step S344), and if it is determined that the dependency is established in consideration of causative / passive, then normally in step S340. Along with the case where it is determined that the dependency is satisfied, the phrase including the word is preferentially processed as a phrase candidate over the difference in the minimum total cost (step S350). Further, the range from the dependent word to the dependent word thus found is registered as the dependent completed range, and a process for managing this is performed (step S360), and it is determined whether or not the dependent search is performed for the entire range ( It proceeds to step S370). In addition, if the establishment of the dependency is denied in consideration of the causative / passive as well as the normal dependency, steps S350 and S36 are performed.
Without 0, the process proceeds to step S370.

Searching for dependency in front of the received word,
The above process (steps S310 to 370) is repeated until all the words except the range registered as searched have been completed, and when the search for the entire range is completed, next, for the plurality of candidates for the received word, , It is determined whether the dependency test is completed (step S
380). In other words, in this example, the candidate "Your company", "Return to office", "Reporter", "Train", etc. for the latter half of "Kisha", which is the catchphrase, are tested to see if there is any relationship in which the dependency relationship is established. Of. A word for which the dependency relation is established is found, and the admissible analysis of the adjunct word is also passed, and the word for which the dependency is established is set to the highest priority as a phrase candidate (step S350). If a dependency relationship is established for multiple candidate words,
The bunsetsu candidates with the highest priority order are registered in the dictionary.

Here, it will be explained that, when a bunsetsu including a word for which dependency has been established is set as a bunsetsu candidate, the bunsetsu including a word for which the bunsetsu does not have the minimum total cost is prioritized. In the example of "Kissha let", the selected phrases "Kishasho" and "Kishashase" are Is the same. But for example,
Assuming that the kana character string "Jokekishakase" was input and the independent word "steam train" existed, as shown in FIG. 9, "steam train" and "return home"
Becomes the path with the minimum cost (solid line G) and becomes the first candidate. On the other hand, the dependency relationship (“reporter
Returning to the office ”)
The first candidate is “above”, “reporter”, or “return to work” that does not become the path with the lowest cost (broken line B in FIG. 9).

FIG. 10 shows the bunsetsu candidates finally obtained by deleting the candidates that cannot be connected by checking the connection between words in the first example sentence "Kissha wa masashi". Therefore, the first candidate for this phrase segmentation is "return reporter to work".

The example of the causative expression has been described above.
As the types thereof, there are the following as notation N1 and verb P. "N1 + P / make" is a causative type of "N1 is + P""N1 makes + P / cause" is a causative type of "N1 + P""N1 + P / cause" is a causative of "N1 + P" Type

In this embodiment, the following causative types that require a particle test over two or more clauses are not allowed. "+ P (adjective verb) to N1""add N1 to + P (adjective verb)""add N1 to + P (adjective)""add N1 to + P (noun)""N1" To + P as in + P (verb end form) "" Make N1 + P "

Further, the processing of the dependency in the passive case will be exemplified. In the case where there is "teaching (study)" as a dependency, the process of writing the inputted kana character string "Seito ga Shieru" is segmented into phrases will be described as an example. FIG. 11 shows a case in which the analysis position advances to “Seito Oshige” and “teaching” is searched backward to find whether a related word exists. There are not only "teaching" as a spoken word but also "teaching" as well as "teaching mathematics" and "teaching", but these are dependent dictionaries 98 It is registered in and you can search for "Kyou" as a headline. Since this search is performed backwards,
The examination starts from the "tail" and reaches "Saints" and "Students", and finds "Students" of "Teach (teaching)" in the dependency with "Teaching" as the headline. Since this dependency is “teaching by students” as it is, it does not pass the admissible analysis of the attached word, and the examination of the dependency is temporarily terminated. After that, when the bunsetsu segmentation test progressed and it reached "Tell me", and the dependency test was made again,
Pass the admissibility analysis of the attached word for the first time by the test of causative / passive dependency. Therefore, it is determined that the dependency is established in the passive case, and "student" and "taught" are the first candidates for the phrase segmentation. This state is shown in FIG.
The first candidate obtained is "student taught". Here, the range from "student" to "taught" is excluded from the subsequent range of dependency search as a range of dependency formation.

The example of the passive expression has been described above.
As the types thereof, there are the following as notation N1 and verb P. "N1 is + P" / "N1 + P" is passive type "N1 is + P" / "N1 is + P" passive type "N1 + P is / P" is "N1 + P" Passive

In the present embodiment, the following passive types which require a particle test over two or more clauses are not allowed. "N1 is (+) + P""N1 is (+) + P""N1 is + P to + N2""N1 is + P""N1 is + P to + N2""N1 to + P (to)" + N2 ”

According to the present embodiment described above, since the dependency information including the causative / passive case is also retrieved at the same time in the process of calculating the cost of words and obtaining the candidates for phrase segmentation, segment segmentation is performed. At the stage of seeking candidates for
It is possible to reflect the information of the dependency including the causative and passive. Since the dependency information is a high-level language activity, it avoids the adverse effect of the uniform selection of phrase segmentation by cost calculation between words and phrases, and selects candidates for segmentation segmentation more in line with the user's intention. It becomes possible to ask. Moreover, since the processing of bunsetsu segmentation by the minimum cost method performed by referring to the independent word dictionary 58 and the adjunct dictionary 68 is performed at the same time as the dependency processing, the processing of segment segmentation using the dependency information is performed in a short time. Can be completed within. When the dependency dictionary 98 is referred to again after the bunsetsu segmentation has been completed, not only the way of dividing the bunsetsu cannot be changed using the dependency information, but also the dictionary is referenced again, so that the processing is performed. It takes time.

Further, when it is determined that the dependency is established once, the range from the received word to the dependency word is set as the dependency establishment range and is excluded from the subsequent search range, so that the dependency range intersects. There is nothing to do. Moreover, it is not judged that two or more dialects receive one related word. Further, since the establishment of the dependency is judged across the adjacent clauses, the dependency test can be correctly performed even if the modifier such as an adverb enters the dependency relation. Therefore, when a plurality of dependencies are satisfied,
As shown in FIG. 3 (A), a combination in which independent dependencies are separately established, or as shown in FIG. 13 (B), a combination in which one word receives two or more dependencies, or FIG.
As shown in (C), a combination in which another dependency is established so as to straddle one dependency is allowed.

Next, a second embodiment of the present invention will be described. The second embodiment uses the same hardware configuration as that of the first embodiment, and its functional blocks are almost the same as those shown in FIG. The functional blocks differ from each other in the structure of the independent word dictionary 58 and the structure of the dependency dictionary 98, that is, the word search process, the dependency test process, and the display process due to the difference in the dictionary structure. These differences and differences in the dictionary structure will be sequentially described according to the processing in the second embodiment.

FIG. 14 is a flow chart showing a kana-kanji conversion processing routine in the second embodiment. This processing routine is a processing which is started when a conversion key (for example, "space key") is pressed after one or a plurality of kana characters are input from the keyboard 24. In addition,
The Kana-Kanji conversion process of FIG. 14 starts when a predetermined number of Kana characters are input or when a delimiter such as “.”, “,” Or “.” Is input even if the conversion key is not operated. It does not matter if it is done. When this process is started, first, a word search process (step S400) and a segmentation process (step S420) are performed. These processes are processes corresponding to the processes of FIG. 3 in the first embodiment.

FIG. 15 shows the details of the word search processing routine. As shown in the figure, when the word search processing routine is activated, first, processing is performed in which the word search start position M is set to the value 1, that is, the head position of the input kana character string (step S402). Next, a process of initializing the variable L indicating the reading length in the word search to the value 1 is performed (step S
404), a process of retrieving this word of reading length L from the independent word dictionary 58 and the auxiliary word dictionary 68 (step S4).
05). Here, the independent word dictionary 58 is composed of a header, an index, and a dictionary body, as shown in FIG. The header is information for managing the dictionary itself. The index and dictionary body are managed by dividing them into basic words, derived words, and meaning examples. The basic word is defined as a word representative of these notations when one word has a derivative notation, for example, when “handling” or “handling” is permitted as the notation for “handling”. It is a word. That is, the basic word only means a word that is typically used in the processing of segmentation and segmentation and dependency processing. Although it is a word recorded in the word dictionary, it is called a representative word, but it does not basically mean a word in the language. Hereinafter, the basic word is referred to as "representative notation" for its display, and as "derivative notation" for derivative words.

The area for the meaning example is an area where the same information as the dependency information described in the first embodiment is managed. Therefore, in the second embodiment, the independent word dictionary 58 and the dependency dictionary 98 are integrated. The information stored in the meaning example area is dependency information centered on the basic word. The relationship between the dependency information and the basic words and the derived words will be described later.

These basic words, derived words, and meaning examples are managed by the B-Tree structure in the dictionary body. The B-Tree structure is a well-known management structure that is adopted when retrieving a large number of data. It is known as a structure in which time is an average time. An example of the B-Tree structure in the dictionary body is shown in FIG. When the B-Tree control block is traced to the word block based on the reading (kana character string), the actual word data is placed here.

Word data such as the basic word area has a data structure roughly shown in FIG. That is, data indicating the data length X of word data is present at the beginning, and then X bytes of actual data follow. The headword length Y is recorded at the beginning of the actual data, and then Y
The entry word of the byte is recorded. The actual word data follows. A word length W is recorded at the beginning of the word data, and immediately after that, a 1-byte flag indicating the presence or absence of kanji data is recorded. Kanji data is recorded after the flag. This Kanji data is composed of a Kanji data length and a Kanji code indicating an actual Kanji character string. After that, word information and part-of-speech information (in some cases, a plurality of pieces of part-of-speech information) are recorded. The word information consists of data indicating the length of the word information and the actual word information.

In this way, whether it is a basic word or a derivative word,
The B-Tree structure can be used to retrieve information about a desired word based on the word heading string. As shown in FIG. 19, these pieces of word information are composed of a separator and data following the separator. As a separator, a display Kanji separator that indicates that the data that follows is Kanji data, or
There is a reading information separator indicating that the data is reading information, a derivative notation separator indicating that the data is a derivative notation, and the like. The display kanji is a record of the kanji corresponding to the representative notation in order to display the kanji by default when one word has a representative notation and a derivative notation. As shown in FIG. 19, in the derivative notation separator, the lower 3 bits of the separator correspond to the number of derivative notations, and the derivative notation 1 and the derivative notation 2 that follow are as shown in FIG.
As illustrated in FIG. 0, the form of transformation to the representative notation is indicated by a number. That is, if the derivative notation 1 is, for example, the number 5, and if the representative notation is “memory”,
Indicates that "with long sound" exists as a derivative notation,
Means "memory". That is, as the derivative notation information, the actual derivative notation itself is not stored, but the form of the derivative notation is stored as a number.
Further, the reading information gives the reading of a word, and gives the reading when the headword is a kanji. This information is used at the time of associative conversion for searching other kanji having the same meaning from the kanji. Note that one basic word and a derivative word (derivative notation) corresponding to this basic word (representative notation) are managed in different areas, but the reading (heading) of the word and the part-of-speech information of the word match. What we do is considered to have a correspondence.

Returning to FIG. 15, after performing a dictionary search (step S405), it is determined whether or not a word matching the reading kana character string of length L from the start position M is found (step S406). When the corresponding word is found, next, a process of reading the data related to the derivative notation among the data attached to the representative word is performed (step S4).
08). If there are various derivative notations for the representative word and the derivative word is stored in the derivative word management area,
It is necessary to replace the ancillary data of the representative word with a derivative notation.
A flag indicating unnecessary is stored. Therefore, it is determined whether or not the flag is referred to and the instruction is made to replace the representative word with the derived word (step S410). When a value indicating the necessity of replacement with a derived word is set in this flag, the representative word searched previously is expanded in the expansion buffer and a mark is added to the word (step S411). The expansion buffer is a storage area for expanding all the representative words and auxiliary words that can form this kana character string into the input kana character string, and is a memory area secured on the RAM 23. .

If the instruction to replace the derived word has not been given, or if the replacement to the derived word has been instructed and the representative word has been marked, the process returns to step S405 and the reading length L It repeats from the process of further searching for the word. When it is determined that the word of reading length L no longer exists in the independent word dictionary 58 (step S4).
06), it is determined whether or not the length of the search word can be extended (step S412). If M + L <A with respect to the total length A of the input kana character string, it is determined that the reading length L can be increased, and the reading is lengthened by one character (that is, L is incremented by a value 1). Processing) is performed (step S414). After the reading length L is increased by the value 1, the processing described above is repeated from step S405.

As a result, the start position M is stored in the expansion buffer.
All reading words from length 1 to the maximum length in are expanded. For word expansion and cost addition,
The procedure is the same as in the first embodiment (see FIG. 4). When the reading length is gradually increased and the word length cannot be extended (step S412), the start position M of the word search is changed to
It is determined whether the input kana character string can be moved toward the end (step S416). If the word can be moved, it is determined that all the search for the word starting from the start position M has been completed, and after the process of incrementing the start position M by 1 is performed (step S418), the reading length L is set to the value. Return to 1 and repeat the above process or go over. Therefore, when these processes are performed,
All the representative words and adjuncts that have the possibility of forming the inputted kana character string are expanded, and the representative words with derivative notations are marked.

With the above processing, the word search processing (step S400 in FIG. 14) is completed. Then, the phrase segmentation processing is next performed (step S420). The phrase segmentation is well known, and the above-mentioned words expanded in the expansion buffer are used to determine the combination of the phrases so that the sum of the values given to each word becomes the smallest value.
There is no particular difference between the phrase segmentation processing and the first embodiment.

Next, a dependency verification process is performed (step S430). The content of the dependency test is the same as that of the first embodiment, but in this embodiment, as described above, only the representative word is expanded in the expansion buffer.
Derived words have not been expanded. In the first embodiment, for example, when there is a dependency of “rule” + “ga” + “change”, if the derivative notation “change” exists for “change”, “rule changes” in the dependency dictionary. And the dependency that "rules change" was remembered. In this embodiment, the dependency test itself is performed only by using the representative notation. Therefore, the dependency test for the character string “Kisoku” + “ga” + “change” is “rule” + “ga”.
+ Only for "change". Therefore, the time required for the dependency check is shortened. In this embodiment, the dependency dictionary is included in the independent word dictionary 58 and is stored in the management area of the meaning example.
FIG. 21 shows an example of the dependency dictionary. The contents of the dependency dictionary are similar to those in the first embodiment (see FIG. 7), that is, reading,
It consists of a related word and an adjunct word that is allowed. In addition, the same rules as those in the first embodiment are applied to the judgment of dependency and the rule of dependency also for causative and passive (see FIG. 13).

In this way, when the dependency test is performed and a word candidate that cannot be determined by the phrase segmentation (for example, “change” and “alternate” in “kisoku ga kawa”) is found, the words for which the dependency is established are determined. If found, processing is performed with this word as the first candidate. After that, the word of the first candidate is replaced with the notation of the word (step S440).
The word replacement process determines whether or not there is a derivative notation for the word selected as the first candidate (for example, “change”),
When there is a derivative notation, the representative notation or derivative notation used first is checked and replaced with that notation. If the word used immediately before is "change," the derivative notation "change" is used in place of the representative word that has been consistently used in the word search, the segmentation process, and the dependency test.

FIG. 22 shows a processing routine for performing the word replacement processing. In this processing routine, the target word to be replaced is first set to the word of the first phrase segmented by the segment segmentation process (step S50).
0), it is determined whether or not the target word is marked (step S510). If the mark is attached, it can be determined that there is a derivative notation and replacement of the derivative notation is instructed. Therefore, the process of replacing the first candidate with the display word is performed thereafter (step S520). The display word is the representative word itself if the representative notation has been used before, and the derivative word if the derivative notation has been previously used and learned. If the target word is not marked, the displayed word is not replaced.

Thereafter, it is judged whether or not there are still unprocessed words (step S530), and if there are any unprocessed words, the target word is moved backward by one (step S540), and the processing is repeated from step S510. If no target word remains, this processing routine ends. In the above-described processing, learning which notation is used as the display word when there is a derivative notation is easy if the word existing at the head of the management area of the derivative word is used as the display word. . In this case, when using a representative word, the representative notation itself (or derivative notation information corresponding to the representative notation) may be stored at the beginning of the derivative word, or the representative notation may be stored at the beginning of the derivative word. A flag indicating whether to use or the derivative notation may be stored.
Further, when using the derivative notation, it is possible to store the notation in the management area of the representative word.

In the display process following the word replacement process,
According to the notation of the replaced word, a process of displaying the word candidate after the kana-kanji conversion on the CRT 26 is performed. In addition,
As for the derivative notation, as shown in FIGS. 19 and 20, the kanji corresponding to the derivative notation is not prepared, but only the number indicating the type of the derivative notation is recorded after the derivative notation separator. Absent. Therefore, in the display processing, according to this number, for example, if the derivative notation information is "4", it is determined that the feed is "permissible" and the representative word "change" is displayed as "change".

According to the present embodiment described above, only the representative word is used as the word candidate until the input kana character string is searched for a word, is divided into words, and is subjected to the dependency processing. Therefore, even when a word for which the derivative notation is recognized is searched, it is not necessary to consider the derivative notation, and each processing can be performed at high speed.
Also, the capacity of the expansion buffer for expanding word candidates can be small. Further, even when there are a plurality of dependencies, there is an advantage that the determination of the dependencies can be easily performed. This will be described with reference to the example of FIG. As illustrated in FIG. 21, “rule” +
There are three types of dependencies: "ga" + "change", "handling" + "ga" + "change", "luggage" + "no" + "handling", and each word is called "changing""handling". Suppose there is a derivative notation. In this case, for example, it is assumed that the notation “rule” + “ga” + “change” is learned for “rule changes”. In this case, if "change" and "change" are managed separately as dependency information,
Even when "changing" is learned during the conversion of "Take care changes" and during the conversion of "Kisoku changes",
"Handling" + "ga" + "change" will be converted. On the other hand, in the present embodiment, learning to use the derivative notation “tolerance” is used for “change” instead of “main rule”, and thus “change” even if another dependency is established. Will be used consistently.

In addition, "there is a different way to deal with the problem"
In the case of conversion of the above, according to the conventional judgment of dependency, even if there is dependency information of "luggage / handling", if there is no dependency of "luggage / handling", "luggage / handling is By comparing the establishment of the dependency chain of “/ change” with the establishment of the chain of “handling / change”, the former dependency will be given priority. As a result, even if "handling / change" is learned immediately before, it will be converted to "package / handling / change". On the other hand, in the present embodiment, since only the derivative notation of “acceptance” is learned for the word “handling”, once “handling” is learned, it is correctly said that “package / handling / change”. Will be converted.

In this embodiment, the word search, the phrase segmentation processing, and the dependency check are performed with the representative word, but a part of these processing is performed using the representative notation and the derivative notation. It doesn't matter.

Although some embodiments of the present invention have been described above, the present invention is not limited to these embodiments. For example, instead of the minimum cost method, another phrase segmentation method such as the two-segment longest matching method is used. It is needless to say that the present invention can be implemented in various modes without departing from the gist of the present invention, such as a configuration using the method described above.

[0086]

As described above, according to the first kana-kanji conversion device and kana-kanji conversion method of the present invention, when the phrase corresponding to the information of the related word and the related word is retrieved, the word attached to the related word Is a permissible adjunct word, and if the adjunct word following the word is a word representing causative or passive, then whether the word attached to the related word is a word corresponding to causative or passive A judgment is made and based on this search result, candidates for phrase segmentation are restricted. Therefore, when the dependency information exists, the dependency is not established simply by the existence of both words, but the dependency is established when the word attached to the dependency allows the relationship between the two terms, In addition, if the received word is a word indicating causative or passive, and if the word attached to the dependent word is a word corresponding to causative or passive, then the candidate candidates are restricted as the establishment of the dependent word. An undesired candidate for segmentation becomes difficult to be selected, and there is an excellent effect that there is a high possibility that desired segmentation will be performed. If the priority of the kanji candidates for each bunsetsu is changed instead of the restriction of the bunsetsu candidates based on the bunsetsu segmentation already estimated by another method, the desired kanji candidate is set as the first candidate. The possibility of being obtained can be increased. Not only changing the priority of word candidates by using dependency information, it is also used for dependency information at the stage of phrase segmentation, and the success of dependency can be determined even in the case of causative or passive. It is possible to obtain the word-spacing candidates that make use of the dependency, which is a higher-level language activity. In addition, it is not necessary to prepare a special dictionary for analyzing causative / passive dependency, and the processing can be performed at high speed.

In addition, according to the kana-kanji conversion device of the second aspect, since only the representative notation is used in the dependency check, it is possible to reduce the time required for the dependency check and the required memory capacity. Play. Furthermore, in the kana-kanji conversion device according to claim 3, the dependency information including causative / passive is used to change the priority of kanji candidates for each phrase instead of the restriction of segmentation and spacing. Using, it is possible to obtain a desired Kanji candidate with high priority. Further, according to the kana-kanji conversion device of the fourth aspect, there is an effect that the search time for separating words is short, and further, the chained dependency is not erroneously selected.

According to the kana-kanji conversion device of the fifth aspect, it is possible to preferentially make a bunsetsu candidate for a word whose dependency is satisfied. In addition, according to the kana-kanji conversion device according to claim 6, it is possible to preferentially select the phrase segmentation based on the dependency and to make the word for which the dependency is satisfied the first candidate for the kana-kanji conversion. .

According to the kana-kanji conversion device of the seventh aspect, it is possible to reduce the amount of information about the attached word. The kana-kanji conversion device according to claim 8 has an effect of being able to find a dependency even when there is a modifier word between the modifier forming the modifier and the modifier. . In addition, when this dependency relation is found, if the range from the starting bunsetsu to the found bunsetsu is excluded from the search range of the next dependency, the dependency crossing It is possible to eliminate erroneous determination of correct dependency and to speed up the dependency search.

According to the kana-kanji conversion device of the ninth aspect, it is possible to select an appropriate word candidate even when no phrase including a word having a dependency relationship is found.

According to the second kana-kanji conversion device and kana-kanji conversion method of the present invention, when the dependency check is performed by referring to the dependency information dictionary, only the representative notation registered in the word dictionary is used for the validation. And specify the words that make up the input character string.When displaying the candidate character strings of this specified word, not only the representative notation stored in the word dictionary but also the derivative notation is used. Since the display is performed as described above, it is possible to achieve both high speed grammar processing including dependency and versatility of notation.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing an implementation form of a kana-kanji conversion function in a kana-kanji conversion device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing hardware for realizing a kana-kanji conversion device as an example.

FIG. 3 is a flowchart showing a phrase segmentation writing process executed in a segmentation segment writing unit 102.

FIG. 4 is an explanatory diagram showing a state of phrase segmentation by the minimum cost method.

FIG. 5 is an explanatory diagram showing another example of phrase segmentation by the minimum cost method.

FIG. 6 is an explanatory diagram showing processing of a dependency check in the embodiment.

FIG. 7 is an explanatory diagram showing an example of a dependency dictionary according to the first embodiment.

FIG. 8 is an explanatory diagram showing an example of a phrase segmentation writing process performed using dependency information.

FIG. 9 is an explanatory diagram showing a state of processing of another sentence example.

FIG. 10 is an explanatory diagram showing priorities of Kana-Kanji conversion candidates in the same case.

FIG. 11 is an explanatory diagram showing a state of processing of dependencies of different types.

FIG. 12 is an explanatory diagram showing the order of priority of kana-kanji conversion candidates in that case.

FIG. 13 is an explanatory diagram showing a type in the case where a plurality of dependencies exist in one input character string.

FIG. 14 is a flowchart showing a kana-kanji conversion processing routine in the second embodiment.

FIG. 15 is a flowchart showing details of the word search process.

16 is an explanatory diagram showing an internal configuration of an independent word dictionary 58. FIG.

FIG. 17 is an explanatory diagram showing how the basic word area is managed.

FIG. 18 is an explanatory diagram showing a structure of word data.

FIG. 19 is an explanatory diagram showing details of word information together with each separator.

FIG. 20 is an explanatory diagram showing an example of derivative notation.

FIG. 21 is an explanatory diagram showing an example of a schematic configuration of a dependency dictionary and representative notation / derivative notation.

FIG. 22 is a flowchart showing a word replacement processing routine.

[Explanation of symbols]

 21 ... CPU 22 ... ROM 23 ... RAM 24 ... Keyboard 25 ... Keyboard interface 26 ... CRT 27 ... CRTC 28 ... Printer 29 ... Printer interface 30 ... Hard disk controller (HDC) 31 ... Bus 32 ... Hard disk 40 ... Character input section 42 ... Conversion Control unit 44 ... Converted character string output unit 50 ... Character string input unit 52 ... Character storage unit 54 ... Independent word candidate creation unit 56 ... Independent word analysis position management unit 58 ... Independent word dictionary 64 ... Adjunct word creation unit 66 ... Adjunct analysis position management unit 68 ... Adjunct dictionary 70 ... Dependency learning unit 70 ... Learning unit 72 ... Independent word learning unit 74 ... Auxiliary word learning unit 76 ... Affix learning unit 78 ... Character conversion learning unit 80 ... Word data creation unit 82 ... Connection verification unit 84 ... Connection verification table 86 ... Cost calculation unit 90 ... Dependency candidate adjustment 92 ... Passive analysis unit 94 ... Particle admissible analysis unit 96 ... Dependency range management unit 98 ... Dependency dictionary 100 ... Word data storage unit 102 ... Phrase segmentation / writing unit 104 ... Dependency transposition information adjustment unit 106 ... Phrase data storage unit 108 ... Converted character string output section

Claims (13)

[Claims] 1. A kana character string is input, the input kana character string is referred to by a word dictionary, and syllabic puncturing is performed to create a syllabic segment slashing candidate. A kana-kanji conversion device that generates a candidate character string to be configured, wherein information on a dependent word and a dependent word that compose a modification of a predetermined bunsetsu, A dependency information dictionary stored together with information, and a phrase provided with the dependency word and a word corresponding to the information of the dependency word when performing the processing of writing the input character string into phrase segmentation And a phrase search means for searching for a phrase having a word corresponding to the information of the related word and the received word, and determining whether the word attached to the related word is the allowable attached word. And the judgment means of When a phrase provided with a word corresponding to the information of the dependent word and the dependent word is searched, when the dependent word following the dependent word is a word indicating causative or passive, the word attached to the dependent word is The system further comprises: second determining means for determining whether the word corresponds to causative or passive; and a phrase candidate limiting means for limiting the phrase segmentation-and-scribbing candidates based on the determination results of the first and second determining means. Kana-Kanji conversion device. 2. The kana-kanji conversion device according to claim 1, wherein the dictionary is a word dictionary storing a reading of a word and a notation corresponding to the reading, and as a notation corresponding to the reading of the word. If there are a plurality of notations, the notation information unit stores the notation determined as the representative notation and the notation determined as the derivative notation, and the phrase searching means is the representative notation stored in the word dictionary. It is a means for searching the dependency using only the phrase, and for the phrase segmentation candidate, using the representative notation and the derivative notation stored in the notation information section of the word dictionary, the converted candidate character string A kana-kanji conversion device having means for displaying candidate character strings. 3. The kana-kanji conversion device according to claim 1, wherein instead of the phrase candidate limiting means, the kanji for each phrase based on the determination results of the first and second determination means. A kana-kanji conversion device equipped with a kanji candidate priority means for changing the priority order of candidates. 4. The kana-kanji conversion device according to claim 1, wherein the bunsetsu searching means sequentially moves forward from the backward bunsetsu as a starting point, excluding the already searched range. A retrograde search means for searching a phrase having a word corresponding to the receiving information, and a phrase having a word corresponding to the dependency information is found by the search, the found phrase from the starting phrase A kana-kanji conversion device provided with already searched range registration means for registering up to the above as a searched range of dependency information. 5. The kana-kanji conversion device according to claim 1, wherein the phrase candidate limiting means is configured such that the first or second determining means determines whether the relevant related word, the dependent word, or the related word is applicable. A kana-kanji conversion device comprising means for preferentially selecting a phrase segmentation containing the word when the existence is determined. 6. The kana-kanji conversion device according to claim 3, wherein, in the kanji candidate prioritizing means, the first or second determining means determines presence of a relevant related word, an adjective and its annex. When determined, a kana-kanji conversion device including means for preferentially selecting a phrase segmentation including the word and selecting the word as a first candidate for kana-kanji conversion. 7. The bunsetsu searching means determines that the dependency relation is established when the auxiliary word existing between the words having the dependency relation has a predetermined specific grammatical structure. The kana-kanji conversion device according to claim 1, further comprising means. 8. The kana-kanji conversion device according to claim 1, wherein the phrase search means refers to the dependency information stored in the dependency dictionary with a predetermined phrase as a starting point, and changes the dependency. When a bunsetsu provided with a word corresponding to information is searched for up to a bunsetsu other than a bunsetsu adjacent to the bunsetsu which is the starting point, and when a dependency relation is found by the bunsetsu bunsetsu searching means, the starting point A kana-kanji conversion device comprising a search range excluding means for excluding the range from the found phrase to the found phrase from the search range by the next alternate phrase searching means. 9. The kana-kanji conversion device according to any one of claims 1 to 3, wherein, in a range in which a phrase including a word having a dependency relation is not found by referring to the dependency dictionary, a combination between words is used. And a kana-kanji conversion device equipped with a means for selecting a combination that maximizes the likelihood of binding between phrases. 10. A kana-kanji conversion for inputting a kana character string, writing out the kana character string that has been input, and writing out a kanji character string, and generating a candidate character string that forms a kana-kanji mixed sentence corresponding to the input kana character string. A device, which is a word dictionary storing a reading of a word and a notation corresponding to the reading, and when there are a plurality of notations corresponding to the reading of the word, a notation defined as a representative notation And a word dictionary having a notation information section that stores the notation determined as a derivative notation, and a dependency information dictionary that stores information on the dependent words and the dependent words that make up the modification of predetermined clauses, A word-spacing processing means for performing phrase segmentation processing using a word dictionary, and for each segment segmented, using the representative notation stored in the word dictionary, the dependency information dictionary Grammar processing means for performing a dependency check to specify a word constituting the input character string, and representative notation and derivative notation stored in the notation information part of the word dictionary for the specified word A kana-kanji conversion device comprising: a candidate character string display means for displaying a candidate character string after conversion. 11. A kana-kanji conversion method for inputting a kana character string, referring to a dictionary, and writing out the input kana character string for each phrase, and generating a kana-kanji mixed character string candidate. When performing the process of segmenting a character string into phrases, refer to the dependency information that stores the dependency information of predetermined phrases together with the information of the admissible adjunct word permitted between the dependency word and the dependency word. When a phrase having a word corresponding to the received information is searched, and a phrase having a word corresponding to the information of the related word and the received word is searched and found, the word attached to the related word is the permissible attachment. If the adjunct word following the received word is a word indicating causative or passive, it is judged whether the word attached to the related word is a word corresponding to causative or passive. , The judgment result of either Kanji conversion method which limits the candidates of the phrase word-separated based on. 12. The kana-kanji conversion method according to claim 11, wherein the kanji candidate priority for each phrase is prioritized based on any one of the determination results, instead of the process of restricting the phrase segmentation candidates. Kana-Kanji conversion method that performs the changing process. 13. A kana-kanji conversion for inputting a kana character string, writing out the kana character string that has been input, and writing out a kanji character string, and generating a candidate character string that forms a kana-kanji mixed sentence corresponding to the input kana character string. A method, wherein, in a word dictionary that stores a reading of a word and a notation corresponding to the reading, as a notation corresponding to the reading of the word, when there are a plurality of notations, a notation determined as a representative notation, The denotation defined as a derivative notation is stored, and the dependency information dictionary stores the information of the dependent word and the dependent word that make up the modification of predetermined phrases, and performs the phrase segmentation processing using the word dictionary. For each of the segmented sentences, only the representative notation stored in the word dictionary is used to perform a dependency test using the dependency information dictionary to construct the input character string. That word to identify, for the word that was said identified using the representative title and derived representation stored in said word dictionary, a kana-kanji conversion method for displaying the candidate character string after conversion.