JPH0855117A - Character processor - Google Patents

Abstract

PURPOSE:To unify retrieve dictionary words having the same notation but different reading while preventing the capacity or processing of a reading substituting dictionary or the like from being increased. CONSTITUTION:Whether character substitution for a character string inputted from an input part 3 is necessary or not is judged based on a character substituting table 7, and when the character substitution is necessary, the inputted character string is saved in a buffer 4, substitution is executed based on the table 7 and retrieval is executed by an independent word dictionary 6 based on a condition stored in a retrieving time restricting condition buffer 9, so that dictionary words having differences in their reading can be unitedly retrieved without preparing a master reading separation type reading dictionary.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character processing device such as a Japanese word processor having a kana-kanji conversion function.

[0002]

2. Description of the Related Art Conventionally, in kana-kanji conversion, a kana heading in a dictionary is searched for with an input character string and the matched word is used. As shown in the block diagram of the kana-kanji conversion device in FIG. 7, the input character string 1 is input from the kana heading of the independent word dictionary 101.
The search means 102 searches for a match with 00, and the result 103 is used to generate a notation candidate. However, as shown in the structure of the independent word dictionary in FIG. 10, words such as “quarrel” and “quarrel”, or “week” and “week” may be written or written in a word. Due to the fluctuation, it was necessary to store the same word as another heading item. However, this not only increases the capacity of the dictionary, but also requires the handling to be done by adding parts of speech and semantic classifications, which is very troublesome in terms of dictionary maintenance.

On the other hand, in the kana-kanji conversion dictionary search mechanism of Japanese Patent Laid-Open No. 4-165557 (published on June 11, 1992), a word group having the same display but different reading has one of the readings as a master. Readings are associated with the notation, and other readings are converted into master readings by a dictionary corresponding to the master readings. This configuration is shown in FIG. As shown in FIG. 11, the dictionary 10 that corresponds to reading the input character string 100
In step 6, the master reading, which is the heading of the independent word dictionary 101, is converted and then searched, so that the separate reading is matched.

However, in the above-described embodiment of the prior art, the dictionary 106 corresponding to the dictionary heading 101 is required for conversion into the master reading, and the same processing as the original dictionary search is performed to search this dictionary. Will be required. Further, since Japanese Patent Laid-Open No. 4-165557 processes the fluctuation of "separate reading of the same notation", fluctuations corresponding to the reading and the notation such as "weak" and "weak" in Katakana are taken into consideration. Absent.

Furthermore, "Genka" for "brothers'fight"
If you read the word as "quarter" and search the dictionary, "prefecture" will become candidates, and the number of useless candidates may increase. Also, if all the readings of "Genka" are replaced with "quarrels", there will be a problem that candidates such as "cost" will disappear.

Further, in the case of converting a plurality of clauses at a time, it may be possible to affect the subsequent retrieval if the reading is left as it is. For example, enter "Kabushiki Gaisha" to search for "company".
Is replaced with "Kaisha", the suffix "ga" or "shigigai" for the conversion candidate "shigiga" → "conductor" becomes the suffix for the conversion candidate "outside the conductor". There is a problem that "outside" etc. cannot be searched.

[0007]

Among the fluctuations in the katakana notation, "be" and "ve", "bo" and "vo", etc. are "best" and "best", "bolt" and "v".
In some cases, another word such as "olt" is originally written. Therefore, in the Kana-Kanji conversion dictionary, the heading "BESTO" should have information about the part of speech and meaning of both words "best" and "best", and "
For the headline, the part of speech of "vest" and the meaning information will be given. In this case, the two dictionary items cannot be unified, and there is no advantage in replacing "be" and "b".

The present invention increases the capacity of the dictionary because it has a plurality of headwords for reading fluctuations, even though the words have the same meaning, such as "Wai" and "Uei". Independent dictionary for conversion by substituting the input character string based on a certain rule without converting to a master reading which is one representative using the reading dictionary and the reading means and searching. By searching for, a plurality of headwords due to reading fluctuations are unified and matched with the dictionary headline.

[0009]

[Means for Solving the Problem] A conversion independent word dictionary provided with information about reading fluctuations, a character replacement table describing replacement conditions and methods, and a character string input from character input means are provided. , Refer to the character replacement table to determine whether there is fluctuation in reading, replace the input character string based on a certain rule, and then match it with the dictionary heading of the conversion independent word dictionary. This makes it possible to effectively search for words in a dictionary that have fluctuations in reading.

[0010]

[Operation] Whether or not replacement of the hiragana character string input from the keyboard or the like is necessary is determined by the character replacement table. If the result of determination is that replacement is necessary, replacement is performed using the character replacement table. When performing replacement, the input character string is temporarily saved in the buffer. After collating with the independent word dictionary for searching, the input character string is returned to the character string saved in the buffer, so that the subsequent conversion is not affected. Further, since the conversion candidate conditions are limited by the character replacement table, it is possible to prevent selection of Kanji candidates for Katakana. Although the replacement is performed in units of kana characters, the replacement such as “kua” → “ka” can also be supported by changing the search start position as shown in FIG.

[0011]

EXAMPLES An example of the present invention will be described in detail below. Needless to say, the present invention is not limited to this embodiment.

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. In this figure, 1 is a control unit,
It contains a grammar table and programs for Kana-Kanji conversion. Reference numeral 2 denotes a display unit including a CRT or LCD, which is used for confirmation of conversion candidates generated from dictionary search results. 3 is an input unit including a keyboard,
Input characters and give instructions for Kana-Kanji conversion. In this embodiment, a hiragana character string is input. Reference numeral 4 is a buffer for storing the converted character string input from the input unit 3. Reference numeral 5 is a pointer indicating from which character in the buffer 4 the dictionary search is to be performed. 6 is an independent word dictionary. Reference numeral 7 is a table for character replacement, which stores conditions such as whether or not character replacement is required, a replacement method, and the like. Reference numeral 8 is a buffer for comparing the previous contents of the buffer 4 and pointer 5 when character replacement is performed, and 9 is a buffer for storing the types of limiting conditions at the time of dictionary search. 3 and 4 both show a configuration example of a table for character replacement. In FIG. 3, character codes to be replaced are collected according to replacement conditions and replacement methods. Conditions, methods, and search conditions are stored for each line. The first line shows that when the first character of the input character string is a kana of kana, it is subject to the replacement process. FIG. 4 shows that the character portion at the pointer position in FIG. 3 is stored by a different method. Among the elements corresponding to the character code, the element corresponding to the character to be replaced is "1". This example also shows an example in which the kana of the dakuon is used as the character to be replaced.

The character string to be replaced is subjected to the processing in the processing method column. Here, the character code is reduced by one. As a result, the voiced sound is replaced by the pure sound. Further, since rendaku is designated as the search condition, only the search condition marked in the part 63 of rendaku among the search conditions in the dictionary of FIG. 5 becomes the search target.
Rendaku means that when two words are compounded, the beginning of the latter word becomes a dakuon. For example, "water"
When the two words "and" are lacking, the result is "water shortage", and the first clear sound in the latter word, "Fu," becomes a dull "bu". .

In the second line of the character replacement table of FIG. 3, when the first character of the input character string is "u" and the second character is "i", "ee" or "ぉ", the replacement processing is performed. This is an example of the target of. As a processing method, the character code is increased by one. As a result, the lowercase letters "i", "e" and "o" are replaced with the uppercase letters "i", "e" and "o".

The fourth line is an example of the replacement process when the first character of the input character string is "ku" and the second character is "a". As a processing method, after replacing "a" with "ka", the search position, that is, the pointer is advanced by one. This process is shown in FIG. By this processing, the character string "kuarutetto" is collated with the dictionary as "karutetto".

FIG. 5 is a schematic diagram showing the contents of the independent word dictionary. Each word is composed of a reading headline 61, a notation 62, and various classification information 63 to 65. Reproducible words such as "fight" and "insufficient" are marked at 63, and words in katakana notation such as "weak" and "weight" are marked at 65. This various classification information is used to limit the conditions when the replacement is performed. For example, in FIG.
In the case of the first line of, that is, when the first character of the input character string is a dakuten, the character string in which the character code is replaced with a clear sound is limited to the rendaku-possible word marked 63 in FIG. Will be searched. For dakuon, it matches both the replaced word and the word before replacement.

For example, when the input character string is "Genka",
Match both "Genka" and "Kenka". However,
For the “quarrel”, only the words that can be rendaku marked with 63 in FIG. 5 are to be matched.

The Katakana word has no notation 62, and the matched input character string is converted to Katakana to generate the notation of the conversion result. As a result, the word "kartet" in the above example is replaced with "kartet", and the search is limited to the katakana word marked 65 in FIG. "Tetto". When collated, the input string "quartet" is converted into katakana, and "quartet" is output.

The replacement of characters by the table of FIG. 3 will be described below with reference to the flow of FIG. (Step S0) The first character of the character string to be searched for in the dictionary, that is, the character in the buffer 4 pointed by the pointer 5 is checked. If "u", go to step S1, if "ku", go to step S1.
If the replacement other than “u” or “ku” is necessary, the process proceeds to each process omitted in the figure. If the replacement is not necessary, the process proceeds to step S5.

(Step S1) "U" pointed by the pointer 5
Find the next character of. "I""e" of the second line of FIG.
If “o”, proceed to step S3. Otherwise, go to step S5.

(Step S2) "ku" indicated by the pointer 5
Find the next character of. If "o" on the third line in FIG. 3 proceeds to step S3, and if "a" on the fourth line in FIG. 3, proceeds to step S4. Otherwise, go to step S5.

(Step S3) The character next to the pointer 5 "i""e""o" is one larger character code "i"
Replace with "e" and "o". The situation before replacement is stored in the buffer 8. It is stored in the buffer 9 that candidates are limited to Katakana. Go to step S5.

(Step S4) The character "a" next to the pointer 5 is replaced with "ka". Pointer 5 is this "ka"
Increase by 1 to point to. The situation before replacement is stored in the buffer 8. It is recorded in the buffer 9 that candidates are limited to Katakana. Go to step S5.

(Step S5) The dictionary 6 is searched with the character string in the buffer 4, starting with the character pointed by the pointer 5.
Go to step S6.

If the input is "weight", step S0
→ S1 → S3 is replaced with "Ueito", and matches "weight" and is stored as a candidate.
However, referring to the contents of buffer 9, "hunger""Uei"
Is not a candidate even if the readings match.

(Step S6) Referring to the buffer 8,
If the buffer 4 and the pointer 5 are changed, the situation before the replacement is restored. Proceed to step S7.

The character string read as "Ueito" here is returned to "Weight". Since the Katakana word "weight" is generated from the character string in the buffer 4, the candidate notation finally displayed on the display unit 2 is "weight".

(Step S7) The character pointed by the pointer 5 is examined. If it is the dull sound in the first line of FIG. 3, the process proceeds to step S8.
Otherwise, the search process ends.

(Step S8) The dull sound pointed by the pointer 5 is replaced with a clear sound having a character code one smaller. The situation before replacement is stored in the buffer 8. It is recorded in the buffer 9 that candidates are limited to words that can be rendaku. Go to step S9.

(Step S9) The dictionary 6 is searched with the character string in the buffer 4, starting with the character pointed by the pointer 5.
Go to step S10.

(Step S10) With reference to the buffer 8, if the buffer 4 has been changed, the condition before replacement is restored, and the search process ends.

The input sentence input from the input unit 3 is stored in the converted character string storage buffer 4, the pointer 5 is set so as to point to the leading character of the converted character string storage buffer 4, and The search process is performed based on the step of 3. The word retrieved by the retrieval process is displayed on the display unit 2.

In the above-mentioned retrieval processing, instead of referring to the buffer 8 to restore the situation before replacement, the clear sound pointed by the pointer 5 may be replaced with the dull sound of the character code one larger. Also, replace the characters in the input string,
Instead of restoring the original character after the dictionary search, as shown in FIG. 7, a character string for search may be extracted from the input character string to replace the character.

[0034]

According to the present invention, by providing means for regularly replacing characters independently of the vocabulary registered in the independent word dictionary, it is possible to unify dictionary words that have fluctuations in reading. Becomes By regularly replacing the characters, it is possible to save an increase in the capacity and processing required by the reading dictionary method, and it is not necessary to prepare a reading dictionary corresponding to a registered word in the independent word dictionary. Also, because it is not limited to words that have the same notation and are read separately,
It is possible to deal with fluctuations in both reading and writing, such as fluctuations in Katakana.

Further, when the text is replaced and the dictionary is searched,
By limiting to words with a certain characteristic described in the dictionary, it is possible to prevent an unnecessary increase of candidates. For example, when "uei" is replaced with "uei", "uei" and the like are not candidates by limiting to katakana only. Furthermore, since the input character string is replaced and the dictionary search is performed, the input character string is returned to the original character, so that it is possible to eliminate the influence on the subsequent search of the dictionary word and the attached word. Further, although the replacement is performed in units of kana characters, the search start position can also be changed, so that the replacement such as “kua” → “ka” can also be supported.

[Brief description of drawings]

FIG. 1 is a diagram showing a kana-kanji conversion device according to an embodiment of the present invention.

FIG. 2 is a block diagram of an embodiment of the present invention.

FIG. 3 is an explanatory diagram showing the contents of a character replacement table according to an embodiment of the present invention.

FIG. 4 is an explanatory diagram showing the contents of a character replacement table according to an embodiment of the present invention.

FIG. 5 is an explanatory diagram showing the contents of an independent word dictionary without separate reading according to an embodiment of the present invention.

FIG. 6 is a flowchart showing a flow of processing of character replacement in an embodiment of the present invention.

FIG. 7 is a block diagram showing another embodiment of recovering a character string before replacement in one embodiment of the present invention.

FIG. 8 is an explanatory diagram showing an example of changing the search start position according to the embodiment of the present invention.

FIG. 9 is a block diagram of a conventional technique.

FIG. 10 is a diagram showing the contents of an independent word dictionary in which another reading of the same word is another item in the related art.

FIG. 11 is a block diagram using a dictionary for replacing separate reading in the prior art.

FIG. 12 is an explanatory diagram of a dictionary that replaces another reading in the related art.

[Explanation of symbols]

 1. Control unit 2. Display unit 3. Input section 4. Converted character string storage buffer 5. Pointer 6. Independent word dictionary 7. Character replacement table 8. Replacement table for replacement 9. Search-only condition buffer

Claims (3)

[Claims] 1. A reading independent fluctuation in a character string input by a character input means, comprising a conversion independent word dictionary to which information about reading fluctuation is added, and a character replacement table describing replacement conditions and methods. Is determined by referring to the character replacement table, the input character string is replaced according to a certain rule, and then the input character string is matched with the dictionary heading of the conversion independent word dictionary. Character processing unit. 2. The character processing device according to claim 1, wherein when the first character of the input character string is a dakuten, the character string in which the character is replaced with a clean sound is collated with the word in the conversion dictionary that can be consecutively confused. 3. The input character string according to a certain rule, when the notation of the pronunciation of katakana varies, the input character string is replaced based on a certain rule and collated with the katakana word in the conversion dictionary. Character processing unit.