JP2806452B2 - Kana-kanji conversion device and method, and recording medium - Google Patents

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 apparatus and method, and a recording medium, and more particularly to a method for quickly and accurately creating a sentence in which Japanese and foreign languages are mixed. Kana-kanji conversion apparatus and method,
And a recording medium.

[0002]

2. Description of the Related Art Recently, with the spread of networks represented by the Internet, electronic mail, net news, and the like have been actively used. In order to utilize such e-mail, net news, etc., it is necessary to create sentences using a computer (word processor). The sentence is increasingly required to include not only Japanese characters but also alphabetic characters and katakana characters of English words.

[0003] In such a continuous kana-kanji conversion algorithm of a conventional kana-kanji conversion system on a computer,
To create a Japanese sentence that includes an alphabetic string of English words, switch the input mode from Japanese (Roman) input mode to fixed alphanumeric input mode and enter alphabetic characters at the time of entering alphabetic characters After that, perform the process of returning the input mode again,
Alternatively, only alphabetic characters are input by another conversion called post-conversion.

[0004]

However, in such a conventional kana-kanji conversion device, it is necessary to switch the input mode or to perform post-conversion, so that it is necessary to input the alphabet characters in the future. It is necessary not only to always be conscious of whether or not to input a character, but also to always be conscious whether or not the current input mode is a mode corresponding to the character that is being input. Furthermore, input mode switching and post-conversion require overhead operations other than those required for original sentence creation, resulting in erroneous input, correction due to it, interruption of thinking during sentence creation, etc. There is a problem that it is easy to do, and it takes time to create a sentence after all.

The present invention has been made in view of such a situation, and aims to quickly and accurately create a sentence in which Japanese and English are mixed.

[0006]

[MEANS FOR SOLVING THE PROBLEMS] The kana kan according to claim 1
The character conversion device receives a character string to be converted.
And the character string received by the receiving means
First analyzing means for analyzing a continuous phrase as a sentence, and receiving means
A string corresponding to a foreign language from the string received by
Extracted by the extracting means and the receiving means
Extracted as a foreign language by the extraction means
The second solution which analyzes the missing part in Japanese
Analysis means, analysis results of the first analysis means, and second analysis means
The analysis result of
Detect the first position where a break matches as a break position
Detecting means, and detecting means for detecting the analysis result of the first analyzing means.
Character string up to the break position detected by
The detection result of the analysis result of the second analysis means is detected by the detection means.
The undetermined character string up to the delimiter position is
And based on the comparison result, the detecting means
Of the undetermined string up to the break position detected by
As an analysis result, the analysis result of the first analysis means and the second analysis
Selection means for selecting one of the analysis results of the means.
And features.

[0007] Kana-Kanji conversion method according to claim 5, varying
Receiving a character string to be replaced, and receiving
The character string received in the step is a continuous phrase as Japanese
A first analyzing step for analyzing and a receiving step
Extract foreign language strings from the extracted strings
Extraction step and the statement received in the receiving step
Whether the string is not extracted as a foreign language in the extraction step
The second analysis method that analyzes the phrase
Step, analysis result of first analysis step and second analysis
Analyze the analysis result of the step toward the end, and
The position where the segment break of the first
And the solution in the first analysis step
Analysis results up to the break position detected in the detection step.
Undetermined character string and analysis result in the second analysis step
Not determined up to the break position detected in the detection step
Are compared based on predetermined selection conditions.
The break position detected in the detection step according to the comparison result
As the result of analyzing the undetermined character string up to
Analysis result in the step and the analysis result in the second analysis step
Selecting one of the fruits .

[0008] The computer readable of claim 9
Kana-Kanji that can convert the input character string
A receiving step for receiving the character string to be converted is provided to the converting device.
And the character string received in the receiving step
First parsing step for parsing clauses as words, receiving
Corresponding to a foreign language from the string received in the step
In the extraction step to extract the character string and the receiving step
Of the received strings, in the extraction step as a foreign language
Segment not analyzed by Japanese
A second analysis step and an analysis result of the first analysis step.
The results and the results of the second analysis step towards the end
Proceed with parsing and find the position where the segment break first matches.
A detecting step of detecting as a break position, and a first analysis
Of the analysis result in the step detected in the detection step
Undetermined character string up to the cutting position and second analysis step
Position detected in the detection step of the analysis result in
Undetermined character strings up to
In the detection step according to the comparison result.
Of the undetermined character string up to the
The analysis result of the first analysis step and the second analysis step
Performs a selection step to select one of the analysis results in the step
A program for causing the program to be recorded .

The selection should be made in the direction with the fewest phrases,
If you have a candidate that includes
, And can be performed as a selection condition.

[0010] Further, a part of the two analysis results is regarded as a next candidate.
Or the uppercase and lowercase letters in the string corresponding to the foreign language,
Differences between square characters and half-width characters can be learned for each candidate or each conversion environment.

[0011] In the kana-kanji conversion apparatus according to the first aspect, the kana-kanji conversion method according to the fifth aspect, and the computer-readable recording medium according to the ninth aspect, the received character string is subjected to continuous phrase analysis as Japanese. In addition, the character string corresponding to the foreign language is extracted from the received character string, and the portion not extracted as the character string corresponding to the foreign language is further subjected to continuous phrase analysis in Japanese. Then, one of the analysis result in Japanese and the analysis result including the character string corresponding to the foreign language is selected based on the selection condition. Therefore, it is possible to quickly and accurately create a sentence in which Japanese and foreign languages are mixed.

[0012]

FIG. 1 is a block diagram showing a configuration example of a kana-kanji conversion device of the present invention. CPU1 is RO
According to the program stored in the M2 or the hard disk drive (HDD) 5, the kana-kanji conversion process is executed. The ROM 2 stores fonts for various characters such as alphabetic characters, numbers, kanji, hiragana, and katakana, in addition to programs. The RAM 3 appropriately stores various data required for performing the kana-kanji conversion process.

The interface (I / F) 4 has an LCD
6 is connected, and the LCD 6 displays the created text. The keyboard 7 is operated when inputting characters.

FIG. 2 shows the configuration of a basic program for kana-kanji conversion processing. When a predetermined character string is input by operating the keyboard 7, the character string is input to the CPU 1 via the interface 4. The CPU 1 supplies this character string to the user input / output unit 21 as character string information before conversion. The user input / output unit 21 converts the pre-conversion character string information into a continuous phrase analysis processing unit 22 and a next candidate analysis processing unit 23.
To supply.

The continuous phrase analysis processing section 22 analyzes the input character string, extracts a character string to be searched in the dictionary, and sends the extracted character string to the dictionary search processing section 24. The dictionary search processing unit 24 searches whether or not the supplied character string is stored in the dictionary 25. Then, the search result is output to the continuous phrase analysis processing unit 22. For example, when a kanji corresponding to a dictionary search character string is supplied from the dictionary search processing unit 24, the continuous phrase analysis processing unit 22 performs a process of replacing the corresponding dictionary search character string with the kanji.

The continuous phrase analysis processing section 22 outputs the analysis result information obtained as a result of analyzing the pre-conversion character string information to the user input / output section 21 as described above. The user input / output unit 21 outputs a converted character string corresponding to the analysis result information supplied from the continuous phrase analysis processing unit 22. CPU1
Causes the LCD 6 to display this character string via the interface 4.

Similarly, the next candidate analysis processing unit 23 extracts a character string to be searched for a dictionary from the pre-conversion character string information input from the user input / output unit 21 and outputs this to the dictionary search processing unit 24. . The dictionary search processing unit 24 searches whether a character corresponding to the input character string to be searched in the dictionary is registered in the dictionary 25, and outputs the search result to the next candidate analysis processing unit 23. The next candidate analysis processing unit 23 puts together a plurality of candidates sent from the dictionary search processing unit 24 as next candidates, and outputs this to the user input / output unit 21.
The user input / output unit 21 converts the character strings of the plurality of next candidates into
The converted character string is output to the CPU 1. CPU1
The character string of the next candidate is also transmitted to the LC via the interface 4.
Output to D6 for display.

As described above, on the LCD 6, a character string obtained by appropriately converting kana (or Roman characters) inputted from the keyboard 7 into kanji is displayed together with the character string of the next candidate. The user selects a character string of a predetermined candidate from the next candidates and creates an intended sentence.

According to the kana-kanji conversion device of the present invention, even when Japanese is mixed with English, a sentence is created by the same operation as in the case of continuously inputting Japanese without changing the input mode. It has been made possible. For this reason, even in the kana-kanji conversion device of the present invention, the program is basically configured as shown in FIG.
More specifically, the program is configured as shown in FIG.

The input unit 41 and the output unit 42 correspond to the user input / output unit 21 in FIG.
Receives a character string input from the keyboard 7 and outputs it to each unit. The output unit 42 receives the character string supplied from each unit and outputs the character string to the LCD 6.

The dictionary search processing unit 43 and the dictionary 44 correspond to the dictionary search processing unit 24 and the dictionary 25 in FIG.
The dictionary search processing unit 43 executes a process of searching for a dictionary search character string supplied from each unit with reference to the dictionary 44 and returning a search result to each unit.

The Japanese-English mixed phrase analysis unit 45 corresponds to the continuous phrase analysis processing unit 22 in FIG. The Japanese-English mixed phrase analysis unit 45 includes a Japanese phrase analysis unit 46 and an English phrase analysis unit 47. As in the case of the conventional kana-kanji conversion device,
The pre-conversion character string is subjected to a continuous phrase analysis process as Japanese, and the English consecutive phrase analysis unit 47 performs a continuous phrase analysis (of English reading) using the pre-conversion character string as English.

The mixed Japanese-English all-candidate analysis unit 48 corresponds to the next candidate analysis processing unit 23 in FIG. 2. The mixed Japanese-English all-candidate analysis unit 48 converts the character string before conversion into Japanese (kana). It has a Roman character kana conversion unit 49 for conversion. The next Japanese candidate processing unit 50 (next candidate analysis means) analyzes Japanese homonyms and the like as the next candidate. The English word reading extraction unit 51 (extraction means) performs a process of extracting English words in order to analyze the character string before conversion as English.

The next English word candidate processing section 52 (next candidate analysis means) analyzes the next English word candidate. The next English word candidate processing unit 52 performs the same processing as the next Japanese candidate processing unit 50 in processing English in English.

The homonym synonym evaluation sort unit 53 evaluates the next candidate obtained as a result of the analysis processing by the Japanese next candidate processing unit 50 and the English word next candidate processing unit 52 according to a predetermined criterion. I do.

The Japanese candidate learning result reflecting unit 54 (learning means) and the English word candidate learning result reflecting unit 55 (learning means)
When a predetermined word is selected from the next candidates of Japanese or English words, a process of learning the result and reflecting the result in the analysis of the next candidate is performed. For example, the English word candidate learning result reflecting unit 55 outputs the candidates having different notations (full-width / half-width,
A process of switching between uppercase / lowercase and alphabet / Katakana according to the immediately preceding learning information (learning information for each candidate, learning information for each conversion environment) is performed.

Next, referring to the flowchart of FIG.
A description will be given of a specific processing example in which, when a predetermined character string is input in alphabetical characters, the character string is converted into kana-kanji characters.

It is assumed that the user operates the keyboard 7 to input, for example, a character string "systemikansitekudasai". This character string is a mixture of Japanese (Roman) and English. In step S1, a process of receiving the pre-conversion information is executed. This processing is performed by the input unit 41 (receiving means). The pre-conversion character string (key input information) received by the input unit 41 is
Japanese reading conversion processing is performed. That is, the input unit 41
The pre-conversion character string received by
6 (first analysis means) and is subjected to continuous phrase analysis as Japanese. As a result, for example, a character string of “Please try sys” is obtained. Then, further in step S3, the Japanese consecutive phrase analysis unit 46 converts the hiragana portion into kanji, and obtains a character string "sys hand not complete".

On the other hand, the input unit 41 also outputs the received pre-conversion character string "systemikansitekudasai" to the English word reading and extracting unit 51 (second analysis means). In step S4, the English word reading and extracting unit 51 performs a process of extracting an English word from the input character string. Details of the English word reading extraction processing are shown in the flowchart of FIG.

First, in step S21, a process of searching the input pre-conversion character string in the English dictionary is performed.
That is, the English word reading extraction unit 51 extracts a character string to be dictionary-searched from the input pre-conversion character string, sends the character string to the dictionary search processing unit 43, and refers to the English dictionary among the dictionaries 44, and Perform a search as a word. The dictionary search processing unit 43 outputs the search result to the English word reading and extracting unit 51.

The English word reading and extracting unit 51 determines in step S22
In, it is determined whether or not there is a searched English word.
In this case, for example, “system” and “site” are searched. At this time, the process proceeds to step S23, and the English word reading extraction unit 51 executes a process of replacing a portion corresponding to the searched character string in the input pre-conversion character string as an English word. In this case, "system" and "site" are processed as English words.

Next, the process proceeds to step S24, in which a word which has not been searched for as an English word is converted into kana. That is, the English word reading extraction unit 51 determines in step S
The character strings “ikan” and “kudasai” that were not searched for as English words in 23 are converted to hiragana. As a result, the pre-conversion character string is eventually converted into “system ikan site please”.

If it is determined in step S22 that the searched English word does not exist, steps S23 and S24 are skipped.

Returning to FIG. 4, assuming that a character string “system ikan site please” is obtained as a result of the English word reading extraction processing as described above, this character string is Is sent to the English consecutive phrase analysis unit 47 (second analysis means). The English continuous phrase analysis unit 47 sends a hiragana portion other than the English word portion of the input character string to the dictionary search processing unit 43, performs a kanji search, and performs a process of appropriately converting it to kanji. Thereby, for example, "Please transfer system site" is obtained.

As described above, in step S3 or step S5, if the continuous phrase analysis as a Japanese reading and the continuous phrase analysis as an English reading are obtained, the process proceeds to step S6 and the subsequent steps. , Is performed.

In step S6, the Japanese / English mixed phrase analyzer 45 determines whether the evaluation up to the beginning of the reading has been completed. In this case, since the evaluation up to the beginning of the reading has not been completed, a NO determination is made, and the process proceeds to step S7 (detection means) . In step S7,
The Japanese-English mixed-phrase analysis unit 45 advances the analysis result until the result of the Japanese-language consecutive- clause analysis result obtained in step S3 matches the phrase segmentation of the English-read continuous- phrase analysis result obtained in step S5. In this example, "siteku
Since "dasai" matches in any of the clauses, this clause is extracted first.

Next, the process proceeds to step S8 (selection means).
It is determined which of the two consecutive phrase analysis results is to be selected with priority. This determination is made according to the following priorities. (1) Priority is given to the phrase having the smaller number of phrases at the phrase break position where the key input information matches. (2) If the number of clauses is the same, the one that contains the candidate used (fixed) immediately before is prioritized. (3) When the above (1) and (2) are the same, priority is given to "Japanese reading".

When the priority order is determined according to this criterion, in this case, the condition (3) is applied,
The phrase "Japanese reading" is selected. As a result, "site
"Please" is selected instead of "Please."

Next, proceeding to step S9, the Japanese-English mixed-phrase analysis unit 45 performs a process of storing the analysis result selected in step S8. In this case, "please" is stored.

Next, the process proceeds to step S10, where both the results of the continuous phrase analysis of the Japanese reading and the results of the continuous phrase analysis of the English reading are respectively advanced to the next analysis processing. Then, the process returns to step S6, and it is determined whether the evaluation has been completed up to the beginning of the reading. In this case, since the evaluation has not yet been completed up to the beginning of the reading, a NO determination is made again, and the process proceeds to step S7. In step S7, "systemik
The phrase "an" will be analyzed. In this case,
The result of the continuous phrase analysis of Japanese reading is "sys hand incomplete", and the result of continuous phrase analysis of English reading is "system transfer".

In step S8, it is determined which of the two analysis results has priority. In this case, “system transfer” is selected in accordance with the criterion (1) that gives priority to the one with a smaller number of clauses.

Next, the process proceeds to step S9, and the process proceeds to step S8.
The analysis result selected in is stored together with the result analyzed in the previous processing. As a result, the analysis result of "Please transfer system" is stored.

Next, the process returns from step S10 to step S6, where it is determined in step S6 that the evaluation has been completed up to the beginning of the reading, so a YES determination is made here and the process proceeds to step S11. In step S11, a process of returning the analysis result stored in step S9 to the output unit 42 is performed. In this case, the converted character string of "Please transfer system" is output to the output unit 42. The output unit 42 outputs this character string to the LCD 6 and displays it.

In the kana-kanji conversion apparatus according to the present invention, a candidate for reading (Japanese reading) in which a character string before conversion is forcibly converted to Japanese (Roman-Kana conversion) and a portion effective as an English word are included. Candidates in the extracted (English reading) extracted and converted into Japanese are processed as homonyms at the same time. This processing will be described with reference to the flowchart in FIG.

For example, in step S 41, “datega” is input to the input unit 41 as the pre-conversion information (key input information).
Shall be received at At this time, in the Japanese consecutive phrase analyzing unit 46, in step S42, the character string "dateg
"a" is forcibly converted to Japanese. As a result, a "date" is obtained.

Next, the process proceeds to step S43, where the next Japanese candidate processing unit 50 performs a next candidate analysis process. Thereby, for example, “Date”, “Tatega”, or “Building” is obtained as the next candidate.

On the other hand, the pre-conversion information "datega" obtained in step S41 is sent to the English word reading extraction unit 51 in step S44, and the English word reading extraction processing is performed.
As a result, the "date" portion is recognized as an English word. As a result, in step S45, a process other than the English word is converted into Japanese by the English consecutive clause analysis unit 47, and "date" is obtained.

[0048] Further, in step S46, the English word next candidate processing unit 52, a homonym obtained in step S45, "date is", "dat e is" or get "dating".

In the embodiment of the present invention, as shown in FIG. 7, lowercase half-width alphabet letters, only the uppercase half-width alphabet letters, all uppercase half-width alphabet letters, and all lowercase half-width alphabet letters, Any of the characters, full-width alphabetic characters in which only the first letter is capitalized, and all-uppercase full-width alphabet characters, full-width katakana characters, or half-width katakana characters are processed as homonyms as the next candidate.

If the next candidate analysis result of Japanese reading or English reading is obtained in step S43 or S46, these next candidates are sent to the homonymous word evaluation sort unit 53, and in step S47. , According to a predetermined criterion (the Japanese candidate learning result reflecting unit 54 described later)
And the learning result of the English word candidate learning result reflecting unit 55)
The evaluation is performed, and a sorting process is performed according to the evaluation result. In step S48, the sorting result obtained as described above is output to the output unit 42, from which it is output and displayed on the LCD 6.

As described above, when a plurality of next candidates are obtained, the user selects a desired one from among them. The result of performing this selection processing is a Japanese candidate learning result reflecting unit 54 and an English word candidate learning result reflecting unit 55
In each, it is reflected. For example, English word candidates with different notations (full-width / half-width, uppercase / lowercase, alphabet /
For katakana, an optimal notation is selected according to the immediately preceding learning information.

In order to realize this processing, information in the dictionary 44 is defined as shown in FIG.

For example, for the word “system”, two kinds of readings, English (alphabetic characters (system)) and Japanese (hiragana (shisutemu)), are registered.
The converted candidates are defined in a format as shown in FIG.

$ E is a flag indicating that the candidate is a special candidate. Next, English (alphabet) candidates are arranged, and then a candidate delimiter flag indicated by @ is arranged. Next, Japanese (Katakana) candidates are arranged.

Further, a special part of speech "English word" (grammatically equivalent to a noun) is adopted as the part of speech.

As shown in FIG. 8, when a conversion operation is performed after "system" is input, the same post-conversion candidate is obtained as when "shisutemu" is input and the conversion operation is performed. Can be
The next English word candidate processing unit 52 and the next Japanese candidate processing unit 50
By converting full-width and half-width characters, or converting lowercase and uppercase characters to English or Japanese candidates after conversion, the next candidates with different notations are automatically created.

The English word candidate learning result reflecting unit 55 and the Japanese candidate learning result reflecting unit 54 perform learning when a candidate is determined. For example, in addition to learning the recent use, the frequency of use, etc., it also learns which notation has been confirmed for each candidate, and determines the type of reading (for example, normal alphabet, initial capital letter,
Learning in the entire conversion environment is performed according to katakana.

Then, when learning is performed, at the time of subsequent conversion, which notation is to be ended is determined based on the next priority. (1) Learning information for the candidate (2) Learning information for the entire environment corresponding to the type of reading (3) Default setting information corresponding to the type of reading

Therefore, for example, if "system" is input,
When “system” was selected last time, next time “syste
m ”is the first candidate. Type "System" and enter "Sys
When “tem” was previously selected, “System” is similarly the first candidate. When “system” is input and “system” is selected last time, “system” is the first candidate next time.

Since the learning is performed in the entire environment, for example, the learning at the time of creating an e-mail and the learning at the time of writing a sentence using a word processor function are performed individually, and the learning in each of them is given priority. Next, learning at the time of executing another function is referred to.

In the above embodiment, English has been described as an example of a foreign language. However, the present invention is also applicable to a case where French, Italian, and other foreign languages are mixed with Japanese. It is possible.

In the above-described embodiment, the program for executing the above-described processing is stored in the ROM 2 or the hard disk drive 5 in advance.
The program may be recorded on a recording medium such as a floppy disk or a CD-ROM disk, and the recording medium may be distributed to users and installed on a hard disk or a RAM. Further, the program may be transmitted via the Internet or another network and installed on a hard disk or a RAM.

[0063]

As described above, according to the kana-kanji conversion device of claim 1, the kana-kanji conversion method of claim 5, and the computer-readable recording medium of claim 9, In addition to analysis results in Japanese, analysis results including character strings as foreign languages are also obtained, and character strings up to the delimitation position of both are compared based on predetermined selection conditions, and corresponding to the comparison results Then, one of the two is selected, so that a sentence in which Japanese and foreign languages are mixed can be quickly and accurately created.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of a kana-kanji conversion device of the present invention.

FIG. 2 is a diagram illustrating a configuration of a program according to the embodiment of FIG. 1;

FIG. 3 is a diagram for explaining a more detailed configuration of a program in the embodiment of FIG. 1;

FIG. 4 is a flowchart illustrating processing of the configuration of the program in FIG. 3;

FIG. 5 is a flowchart illustrating a more detailed process of the English word reading extraction process in step S4 of FIG. 4;

FIG. 6 is a flowchart illustrating a process of a Japanese-English mixed all candidate analysis unit in FIG. 3;

FIG. 7 is a diagram illustrating the configuration of a homonym.

FIG. 8 is a diagram illustrating the definition of information in a dictionary.

FIG. 9 is a diagram illustrating a special format of a candidate after conversion.

[Explanation of symbols]

 Reference Signs List 21 User input / output unit 22 Continuous phrase analysis processing unit 23 Next candidate analysis processing unit 24 Dictionary search processing unit 25 Dictionary 41 Input unit 42 Output unit 43 Dictionary search processing unit 44 Dictionary 45 Japanese-English mixed continuous phrase analysis unit 46 Japanese consecutive phrase analysis unit 47 English phrase and phrase analysis unit 48 Japanese-English mixed candidate analysis unit 49 Romaji kana conversion unit 50 Japanese next candidate processing unit 51 English word reading and extraction unit 52 English word next candidate processing unit 53 Homonymic synonym evaluation sort unit 54 Japanese candidate learning result Reflection unit 55 English word candidate learning result reflection unit

Continuation of front page (56) References JP-A-5-282360 (JP, A) JP-A-63-316162 (JP, A) JP-A-64-53265 (JP, A) JP-A-4-75162 (JP, A) JP-A 8-106461 (JP, A) JP-A 64-18872 (JP, A) JP-A 8-212216 (JP, A) JP-A 5-342192 (JP, A) 60-20234 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G06F 17/21 G06F 17/22

Claims (12)

(57) [Claims] A receiving means for receiving a character string to be converted; a first analyzing means for performing continuous phrase analysis on the character string received by the receiving means as Japanese; and a foreign language from the character string received by the receiving means. Extracting means for extracting a character string corresponding to a word; and second analyzing means for analyzing a portion of the character string received by the receiving means, which is not extracted as a foreign language by the extracting means, as a continuous phrase analysis in Japanese Detecting means for analyzing the analysis result of the first analysis means and the analysis result of the second analysis means toward an end, and detecting a position where a segment break of the phrase first matches as a break position; An undetermined character string of the analysis result of the first analysis means up to the break position detected by the detection means, and an analysis result of the second analysis means. The undetermined character string up to the break position detected by the detection means is compared based on a predetermined selection condition, and the undetermined character string up to the break position detected by the detection means is compared with the comparison result. A kana-kanji conversion device comprising: a selection unit that selects one of an analysis result of the first analysis unit and an analysis result of the second analysis unit as an analysis result of a character string. 2. The method according to claim 1, wherein the selecting means selects the one having the smaller number of phrases.
As the first selection condition, the candidate used immediately before is included
Is used as the next selection condition, and the first analysis means
The analysis result is used as the next selection condition.
2. The kana-kanji conversion device according to claim 1, wherein the selection is performed . 3. The foreign language extracted by the extracting means.
At least uppercase and lowercase letters, full-width characters of the corresponding string
Letters and single-byte characters, or only capital letters and small letters thereafter
A string that is a character and a string that starts and ends with a lowercase letter
3. The kana-kanji conversion device according to claim 1 or 2, further comprising a processing unit for processing as a next candidate . 4. Learning the notation finally selected by the user
The kana-kanji conversion device according to claim 1, 2, or 3, further comprising a learning means for reflecting the result in a next process . 5. A kana-kanji conversion for converting an input character string.
In the kana-kanji conversion method in the conversion device, a receiving step of receiving a character string to be converted , and converting the character string received in the receiving step into Japanese
A first parsing step of parsing a phrase as a sequence, and a foreign language from the character string received in the receiving step.
Extracting a character string corresponding to the character string, and among the character strings received in the receiving step,
Part not extracted as a foreign language in the extraction step
Second parsing step for analyzing multiple phrases in Japanese
And the analysis result of the first analysis step and the second analysis step.
Analyze the analysis results of the step toward the end,
The position where the break of the clause first matches as the break position
A detecting step of detecting, and the detecting step of the analysis result in the first analyzing step.
Undetermined character string up to the break position detected in the
Detecting the analysis result in the second analysis step;
The undetermined character string up to the break position detected in the
Comparing based on predetermined selection conditions and corresponding to the comparison result
To the break position detected in the detection step.
As the analysis result of the undetermined character string, the first analysis step
Analysis results in the second analysis step and the analysis results in the second analysis step.
Kana-Kanji conversion method characterized by comprising a selection step of selecting one of the fruit. 6. In the selecting step, the number of phrases is small.
Is the first selection condition, and the candidate used immediately before is included.
The first analysis is defined as the next selection condition.
The analysis result in the step is further
The kana-kanji conversion method according to claim 5 , wherein the selection is performed as a matter . 7. The foreign language extracted in the extracting step
At least uppercase and lowercase letters, full-width characters of the corresponding string
Letters and single-byte characters, or only capital letters and small letters thereafter
A string that is a character and a string that starts and ends with a lowercase letter
7. The kana-kanji conversion method according to claim 5 , further comprising a processing step of processing as a next candidate . 8. Learning the notation finally selected by the user
The kana-kanji conversion method according to claim 5, 6 or 7, further comprising a learning step for reflecting the result in a next process. 9. A kana-kanji conversion for converting an input character string.
A receiving step of receiving a character string to be converted , and converting the character string received in the receiving step into Japanese
A first parsing step of parsing a phrase as a sequence, and a foreign language from the character string received in the receiving step.
Extracting a character string corresponding to the character string, and among the character strings received in the receiving step,
Part not extracted as a foreign language in the extraction step
Second parsing step for analyzing multiple phrases in Japanese
And the analysis result of the first analysis step and the second analysis step.
Analyze the analysis results of the step toward the end,
The position where the break of the clause first matches as the break position
A detecting step of detecting, and the detecting step of the analysis result in the first analyzing step.
Undetermined character string up to the break position detected in the
Detecting the analysis result in the second analysis step;
The undetermined character string up to the break position detected in the
Comparing based on predetermined selection conditions and corresponding to the comparison result
To the break position detected in the detection step.
As the analysis result of the undetermined character string, the first analysis step
Analysis results in the second analysis step and the analysis results in the second analysis step.
To perform a selection step to select one of the results.
Computer reading characterized by recording the program
A removable recording medium . 10. In the selecting step, the number of phrases is small.
Is the first selection condition, and the candidate used immediately before
The included condition is used as the next selection condition, and the first solution
Analysis result in the analysis step is further selected
10. The method according to claim 9 , wherein the selection is performed as a condition.
A recording medium according to claim 1. 11. The foreign language extracted in the extraction step
, At least uppercase and lowercase, all
Square and half-width characters, or only the first letter is uppercase and
Strings that are lowercase and characters that start and end with lowercase
Perform further processing steps to process the column as the next candidate
The recording medium according to claim 9, wherein a program for causing the recording medium to be recorded is recorded. 12. Learning the notation finally selected by the user
And further execute the learning step to be reflected in the next processing
12. The recording medium according to claim 9, wherein a program for causing the recording medium to be recorded is recorded.