JPH0833886B2 - Kana-Kanji conversion device and method - Google Patents

Description

[Detailed Description of the Invention] A. Field of Industrial Application The present invention relates to a kana-kanji conversion device for converting a Japanese sentence of a continuous phrase input by kana into a kana-kanji mixed sentence, and particularly possible for kana-kanji conversion after re-segmentation. It is designed to obtain accurate conversion results.

B. Prior art Kana-Kanji conversion is known, which uses a Japanese text input device as a Japanese input device to input a kana character string and convert it into a kana-kanji mixed sentence consisting of multiple clauses. After conversion, it may be necessary to change the position of the segment break. In order to solve this problem, for example, as shown in Japanese Patent Laid-Open No. 59-176820, once the entire kana-kanji mixed sentence after conversion is once returned to the kana character string, and the bunsetsu delimiter position is specified by manual operation, A method of performing successive conversion again is known.

In this case, the time required for the re-conversion becomes the same as that for the conversion in the single phrase, and the advantage of using the consecutive phrase kana-kanji conversion method is lost.

On the other hand, if you do not explicitly specify the position of the phrase break,
Even if the re-conversion instruction is given, since the Kana-Kanji conversion in the normal time is applied to the conversion target area, the same punctuation position error as before the re-conversion instruction often occurs, and in order to obtain the correct conversion result. It takes a lot of work.

This is for the following reason. When a range containing multiple bunsetsu is specified as the target of re-conversion, the specified part often contains an error in bunsetsu delimiter. This is because the same mistakes are made again by applying the same conversion method that made a mistake.

C. Problems to be Solved by the Invention This invention was made in consideration of the above circumstances, and is a kana-kanji conversion that requires re-segmentation.
The purpose is to provide a kana-kanji conversion device that can quickly obtain accurate conversion results.

D. Means for Solving Problems In the present invention, in order to achieve the above object, different conversion methods are used for normal conversion and reconversion.

For example, in normal conversion, conversion is performed based on the score (penalty) of each word, and candidates are output in order from the one with the smallest number of phrases at the time of reconversion that requires segmentation again.

E. Embodiment An embodiment in which the present invention is applied to a kana-kanji conversion device for consecutive phrases will be described below.

FIG. 1 is a block diagram of a kana-kanji conversion device according to an embodiment of the present invention. In FIG. 1, the keyboard 1 is
In addition to the character keys for inputting the characters to be converted to kana-kanji, the conversion key for converting the input kana-character string into a kana-kanji mixed sentence, the cursor key for moving the cursor, and various proofreading edits. Function keys etc. are provided.

The first conversion processing unit 2 performs continuous sentence kana-kanji conversion for input from the keyboard 1. At this time, the dictionary 5 is searched according to the input kana character string, and first, one or more feasible phrase sequences (continuous phrase conversion candidates) are obtained. An evaluation is performed using a score (penalty) attached to a word chain. A score associated with a word is, for example, frequency information of an individual word, and a score associated with a word chain is, for example, grammatical information. The optimum conversion candidate is selected based on the above scores. Note that the conversion method at this time may be a method such as the longest match method or the two-segment longest match method.

The second conversion processing unit 3 performs continuous sentence-kana-kanji conversion so that the number of clauses in the kana-kanji mixed sentence obtained for the portion for which the reconversion instruction is given is minimized. As a conversion method at the time of re-conversion, it is possible to add a peculiar score of a word in the dictionary 5 to the minimum clause number method.

The character string buffer 4 stores a character string input from the keyboard and is used as an input to the conversion processing units 2 and 3.

The conversion method switching unit 6 determines whether or not the reconversion instruction is the selection of a homophone, for example, by determining whether or not the portion for which the reconversion instruction is given matches one word base or one phrase, and selects the homophone. If not, the contents of the character string buffer are given to the second conversion processing unit 3 and control is passed to the second conversion processing unit 3. (FIG. 2) In the display 7, not only the character string input from the keyboard is displayed, but also the character string resulting from the kana-kanji conversion is displayed.

Next, the conversion result and the correction process for the input kana character string "Uranii Umugen Soha Tsuke" are shown. In the kana-kanji conversion in the normal mode used in the initial conversion, for example, it is assumed that the conversion result as shown in FIG. 3 is obtained. Some of the conversion result candidates at this time are shown in FIG. For this conversion result, the operator designates the range to be reconverted as shown in FIG. 4 (a) and gives an instruction for reconversion. The conversion method switching unit 6 judges that this operation is an instruction for reconversion and is not an operation for selecting a homonym synonym, and a portion corresponding to the portion in the character string buffer for which reconversion is instructed. Is converted in the re-conversion mode and the conversion result is output to the display 7. An example of the conversion result at this time is shown in FIG. It should be noted that the designation of the phrase recut range in FIG.
Although I am working on the converted kana-kanji mixed sentence, it may be done after returning to kana once, which is also the usual way.

Further, FIG. 6 shows an operation by the minimum clause number method, taking a case where the portion "Urani Umugen no Hatsutsu" is set as a clause recutting range. In this case, "discovery of uranium element" is output. In the above description, two conversion processing units are provided on the hardware, but one processing unit can also be used as two conversion processing units by software.

F. Effect As described above, the present invention uses two or more kana-kanji conversion processing units for kana-kanji conversion of a kana-character string consisting of a plurality of phrases, and performs initial conversion for a conversion result including a bunsetsu boundary error. By performing re-conversion using a conversion method different from that described above, it is possible to correct the result of the Kana-Kanji conversion that includes a bunsetsu boundary error with less labor than when the same conversion method is used during re-conversion. As a result, it is possible to significantly reduce the time required for Japanese input including the correction operation.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the kana-kanji conversion device of the present invention, FIG. 2 is the same flow chart, FIG. 3 is a description of conversion in the normal mode, and FIG. 4 is a description of conversion during re-conversion.
FIG. 5 is a part of the conversion result table in the normal mode,
The figure shows a part of the conversion result table in the reconversion mode. 2 ... 1st conversion processing part, 3 ... 2nd conversion processing part, 5 ... conversion method switching part.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP 62-143175 (JP, A) JP 61-198361 (JP, A) JP 60-20234 (JP, A)

Claims (5)

[Claims] [Claim 1] The first Japanese sentence of a continuous phrase input with Kana
In accordance with the first conversion means for converting into a kana-kanji mixed sentence according to the standard, and the second conversion means for converting a Japanese sentence of a continuous phrase input in kana into a kana-kanji mixed sentence according to the second criterion, Control for selectively operating the first conversion means and the second conversion means for converting the whole Japanese sentence input by the above kana as a conversion candidate having a higher rank for a phrase sequence having a smaller number of converted phrases. A kana-kanji conversion device characterized in that the conversion means executes the conversion by the first conversion means under normal conditions, and the conversion is executed by the second conversion means immediately after re-segmentation. 2. The kana-kanji conversion device according to claim 1, wherein the first conversion means evaluates the conversion candidates based on the scores of the words forming the phrase sequence and the scores of the chain between the words. 3. The kana-kanji conversion device according to claim 1, wherein the first conversion means executes the conversion based on the longest match method. 4. The kana-kanji conversion device according to claim 1, wherein the first conversion means executes conversion based on the two-segment longest-match method. 5. A Japanese sentence of a continuous phrase input by kana is the first sentence.
Based on the standard of kana and kanji, a step of accepting an instruction to rephrase the phrase, and based on the received instruction to rephrase the phrase, the whole sentence of Japanese input in kana is converted and converted. A kana-kanji conversion method comprising the step of converting to a kana-kanji mixed sentence according to a second criterion in which a phrase sequence having a smaller number of subsequent phrases is a conversion candidate having a higher rank.