JPS61255462A - Kana kanji converting device - Google Patents

Abstract

PURPOSE:To execute easier a selecting operation by a user by designating a scope in a sentence with another exclusive-use key and decreasing the candidate of a homophone outputted when the word in the special category is inputted by a proper noun. CONSTITUTION:A KANA (Japanese syllabary)/KANJI (Chinese character) converting part 2 is composed of, for example, a converting control part 2a, a clause analyzing part 2b, a word retrieving part 2c, a grammar table 2d and converting dictionaries 3 and 4, and for the first character string transferred from an input device 1, the second character string composed of the KANJI- mixed displaying character string equivalent to the string is obtained. For the usual input, the first dictionary 3 only is investigated, and by a proper noun key 1 as a special key in the input device 1, the independent word is calculated for the second dictionary 4 only concerning the part to receive the indication in an input character system.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a kana-kanji conversion device that can convert solid input into kanji and kanji in a Japanese word processor or the like.

[Technical background of the invention and its problems]

In recent years, advances in kana-kanji conversion technology for Japanese word processors have led to the adoption of methods that convert long units such as words and single phrases to compound words and compound phrases with a single conversion instruction. This method can combine and output multiple words such as a prefix and a noun, a noun and a noun, a noun and a suffix, etc., or output a series of multiple phrases. The operability is improved compared to the conventional method of specifying conversion while keeping in mind the units of .

However, while not having to specify the units of phrases or words, while improving the input operability, it also means that the degree of freedom of interpretation as Japanese is increased to an extreme extent in some cases. In order to select a word or phrase, it may be necessary to press the key many times to specify the next candidate from the homophone data. Also, it is necessary to operate while maintaining these many homophone candidate data internally. has limitations in terms of internal memory capacity and increased processing time, and generally limits the number of interpretations and performs kana-kanji conversion, and may sometimes fail to output the desired phrase.

Therefore, it often happens that the word is not converted into the intended word and is output as a different phrase sequence or as a composite form, and in order to correct these correctly, input the incorrectly converted part in kana again and convert It was necessary to give instructions.For example, if you input a sentence in kana such as ``Today, I'm going to Koji and Kaimono,'' and then convert it into a sentence mixed with kanji, you can write ``Today I'm going to do some construction work and go shopping.'' There are many interpretations, such as ``I'm going to sweep the streets today or go to the foundry.'' Since the Kana-Kanji conversion mechanism only supports grammatical analysis, unnatural Japanese sentences are actually generated as homophone data. In order to switch the result displayed as ``I'm going shopping with Koji today'' to the homophone of the desired sentence, a considerable amount of next candidate operation is required. In this example, especially when there are 10 or 20 different homophones, such as the proper noun "Koji", the number of combinations of homophones increases, making it difficult for the user to correct them. It had the disadvantage of being.

[Purpose of the invention]

The purpose of the present invention is that when inputting a word in a specific range, such as a proper noun, the range is specified in the sentence using a separate key, thereby reducing the number of homophone candidates that are generated. An object of the present invention is to provide a kana-kanji conversion device that makes selection operations by a user easier.

[Summary of the invention]

When inputting a sentence using solid text input, the present invention allows for words in a specific category such as proper nouns to be input by specifying the range using a specific key. The Kana-Kanji conversion result is output as homophone data consisting only of words in the range, and for ranges outside the specified range, the result is output as homophone data consisting only of words that do not include the words in that range.

〔Effect of the invention〕

According to the present invention, when performing kana-kanji conversion using solid input, for example, by specifying only the proper noun part, it is possible to convert ordinary words such as ``koji'' and ``koji'' into common words such as ``koji'' and ``koji.'' When people's names are mixed together, homophones become more pronounced,
In addition, it is possible to reduce the discrepancy between interpretations such as ``Today is with Koji'' and ``Today is with Koji'', which has a great practical advantage for the user in that it improves the operability to obtain the desired sentence. is played.

[Embodiments of the invention]

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

FIG. 1 is a schematic configuration diagram of an embodiment device. The input device 1 includes a keyboard device, a voice recognition device, and a kana character reading device. A character string (first character string) representing the pronunciation inputted through the input device 1 is converted into, for example, a kana character code, and then is provided to the kana-kanji converter 2 .

Note that the first character string representing the above-mentioned reading is shown as 1, for example, hiragana, 1 kishikana, λ, romaji, etc.

The kana-kanji conversion unit 2 is composed of 1, for example, a conversion control unit 2a, a phrase analysis unit 2b, a word search unit 2C, a grammar table 2d1, and conversion dictionaries 3 and 4, and an input device.
For the first character string transferred, a second character string consisting of a corresponding display character string mixed with kanji is obtained. These data are temporarily stored in the phrase series memory 21.

The first conversion unit 3 contains general words, such as general nouns, verbs, and adjectives, and is composed of pronunciation, kanji notation, and part-of-speech classification information, as shown in Figure 2. This is table data. Further, the second conversion dictionary section 4 separately stores words and phrases in a specific predetermined category, and is, for example, a table in a similar format of proper nouns as shown in FIG. For the pronunciation string transferred from the input device 7, the conversion control unit 2a detects special character codes such as punctuation marks and line feed instruction codes in the input data, and associates the previous kana character strings as a series of phrases. , performs kana-kanji conversion. As this method, for example, the longest match method, the longest two clause match method, the minimum number of clauses method, etc. have been tried in the past.

For example, when using the minimum clause number method, the conversion control unit 2a instructs the clause analysis unit 2b to extract as many parts of the input character string that can be considered as clauses, and then It combines clauses to create a sequence corresponding to the entire input string, and among them, selects and outputs the sequence with the smallest number of clauses. For example, in response to the input "The weather is good", (7) "The weather is/Good J, &)
Various sentence m sequences can be obtained, such as "r weather/painting/good" (y) "point/feel/good", but this has the least number of clauses (7)
is selected and outputs a sentence containing kanji. (/ indicates a break between phrases) Note that the phrase analysis unit 2b analyzes words,
The search unit 2C is instructed to match independent words in the input character string using the conversion dictionary unit, and the continuation of the obtained independent word part is tested for conjugated endings using the grammar table 2d. , and test the attached word string. For example, in the case of the clause ``weather'', the independent word part is ``weather''', and
The ``1'' that follows the noun heavy oil is interpreted as a case particle.

Here, the conversion dictionary includes a first dictionary 3 and a second dictionary 4, and only the first dictionary is checked for normal input. By using the proper noun key 1 as a specific key in the input device 1, an independent word is assigned to the second dictionary 4 only for the designated part of the input character sequence.

The thus obtained kanji-mixed sentence is transferred to the output data memory 6 in the output control section 5, and then to the display device 7.
is output to. The input character string is also output to this display device 7 until the kana-kanji conversion result is output.

Next, an example of the operation of the kana-kanji converter 2 will be described using a specific input example shown in FIG. 4 and a diagram 70 of the conversion process shown in FIG.

Input device? The input character string transferred from 1i1 is sent to the output control unit 5 via the conversion control unit 2, and is first output as a kana character string on one display screen.

(Fig. 6 61, 62, 63) 7a As shown in ``Kiyouha'' is input, and then key 1 is pressed to indicate the proper noun.
When a is pressed, a one-color " symbol, for example, will be displayed as shown in 7b. Next, enter the person's name "Koji".
By pressing the key 1a again, the end of one proper noun part is displayed as "time".

(7C) The display of this symbol is such that, for example, each time the key 1a is input, the symbol is automatically switched alternately to the symbol of the start point and the end point. (Fig. 6 64,65,66゜67) Next, as shown in 7d, the input to the end of the sentence is completed, and the period "0" is entered.
'' is input, the kana-kanji conversion unit 2 extracts units of phrases and creates a series of phrases. (Fig. 6 68.69
．． 72.73) For ``0 Kiyoha'', search the first conversion dictionary 3 to extract the independent word ``Today'', refer to the grammar table 2d, connect the particle ``wa'', and change it to ``° Today, we will recognize the phrase ``.Next, for the part ``(Koji)'', search the second conversion dictionary 4 according to the proper noun instructions, for example, ``Koji'', ``Shinji'', etc. He picks out some homophones, such as "Koji."

(Fig. 6 77, 78.79) Next, 6 and so on.
. . ”. At this time, the preceding character string may be regarded as an adjunct to the proper noun word “Koji,” or it may be regarded as the beginning of a general phrase that follows.

In the former, by referring to the grammar tape/I/2d for the adjunct, the phrase "Koji to ga" or 6 Koji to is recognized. In the latter, the first conversion dictionary 3 and the grammar tape!L are used.
By searching for /2d, phrases such as ``submit'', ``toka'', and ``to'' are recognized. is used to find the phrase string up to the end of the sentence, for example, as shown in Figure 5. (Figure 6 74.75'') These operations are carried out from the conversion control unit 2a to the phrase analysis unit 2b and word search unit 2C. By giving an instruction to start the operation and transferring the results obtained in the reverse order, a possible series of clause data is stored in the clause series memory 21 in the 1 conversion control unit 2a. (Fig. 6 7
6) Next, if we look at the rule that prioritizes the minimum number of clauses, for example, Figures (1) and (2) are selected as the results, and (6)
Figure 70) Among these, for example, make a decision such as giving priority to the one with one particle attached, and use the result of (2) as the first
The ranking is displayed and output as shown in FIG. 4 78. (Fig. 6, 71) The squares in Fig. 5 indicate the representative words (for example, the first rank in the conversion dictionary) of other homophone words. Here, it is displayed with an overline. When the operator does not find the desired word or phrase, he or she can select the desired word from homophone candidates by performing operations such as "next candidate", just like in a conventional word processor.

Note that the present invention is not limited to the above embodiments.

For example, in response to an operator's instruction, the conversion dictionary section is
Although the configuration is shown in which the dictionary 3 and the second dictionary 4 are separated, they may be combined into one dictionary data structure. At this time, following the operator's instructions, you can search for a specific category of words using the part-of-speech information of each word as clues. For example, it becomes possible to specify the field of use.

In addition, the present invention can be implemented with various modifications without departing from the gist thereof.

[Brief explanation of drawings]

FIG. 1 is a schematic configuration diagram of an apparatus showing one embodiment of the present invention, and FIG.
Figure 3 shows an example of the first conversion dictionary, Figure 3 shows an example of the second conversion dictionary, Figure 4 shows an example of display output, and Figure 5 shows an example obtained as a phrase series. 6 are diagrams showing the flow of conversion processing in the embodiment apparatus. l...Input device l...Unique multi-frame instruction key 2...
Kana-kanji conversion unit 2a... Conversion control unit 21... Clause series memory 2b... Clause analysis unit 2C... Clause analysis unit 2C... Word search unit 2d... Grammar table 3.
... Conversion first dictionary 4 ... Conversion second dictionary 5 ... Output control section 6 ... Output memory data 7 ... Display device 7
a~7e... Display output example Agent Patent attorney Noriyuki Chika (and 1 other person) Figure 3 Figure 4 Figure 5

Claims (1)

[Claims] an input means for inputting a kana character string and a specific word designation signal; a first dictionary storing kana characters and corresponding kanji; and a first dictionary storing kana characters and corresponding kanji in a specific field such as proper nouns; a second dictionary stored therein; a clause analysis means for extracting clauses by grammatically analyzing the kana character string inputted from the input means; and a clause analysis means for extracting clauses from the first dictionary according to the clauses extracted by the clause analysis means. If a kana character included in the phrase is input together with a specific word designation signal, a search means for searching the corresponding kanji from the second dictionary, and a kanji searched by the search means are provided. A kana-kanji conversion device characterized in that it is equipped with a conversion means for converting the input kana character string into a kana-kanji mixed sentence of a continuous clause, and an output means for outputting the kana-kanji mixed sentence converted by the conversion means.