JPH0326868B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to a kana-kanji conversion device that converts a Japanese sentence input in kana code into a sentence containing kana-kanji and outputs the converted sentence.

(Prior art) In conventional kana-kanji conversion devices, a kana character string is decomposed into all possible elements, combined, and a combination test is performed to obtain a correct kana kanji character string, or a kana character string is Each time the words were shortened from the back, the adjuncts were broken down and a combination test was performed.

For this reason, there are disadvantages in that the area in which the adjunct words are expanded becomes large and processing time is required.

(Purpose) In view of these problems, the present invention provides an attached word certification conversion device within a kana-kanji conversion device.
The purpose is to reduce the dependent word development area, reduce the dependent word verification time and the dependent word character development time, and perform faster kana-kanji conversion processing.

(Example) FIG. 1 is a diagram showing a device to which the kana-kanji conversion device of the present invention is applied. Reference number 1 in the figure is based on data input from the input unit (keyboard).
This is the central processing unit that controls the entire system. The character string and control code written by the input section 2 are sent to the kana-kanji conversion device 8, where they are converted into kana-kanji, outputted as candidate data to the kana-kanji candidate selection section 3, and displayed on the display section 4. . Here, according to instructions from the operator, the functions of the kana-kanji candidate selection section 3 are utilized to display candidates one after another on the display section 4, and an appropriate one is selected and outputted to the output section 5.

Further, the kana-kanji conversion device 8 inputs the kana character string to be converted into the grammar analysis unit 1 through the kana character string input unit 9.
Send to 0. Grammar analysis unit 10 (enumerates adjunct words by means of assuming that the independent word stem is a certain range of kana character strings, and by means of testing the possibility of using the cut out independent words and the rest of the kana character strings as an enumeration of adjunct words. The part of speech of the kana adjunct character string that has been verified to be conjugated with the first adjunct is tested, and if the combination is possible, the adjunct character strings are combined and sent to the display section 4. Based on the assumption that the data is always in phrase units, here we assume whether it is a phrase with a prefix or suffix or not, specify the independent word length, and cut out the independent words. The remaining kana character strings from which the independent words have been cut out can be assumed to be adjunct words, so they are sent to the adjunct word verification section 14. Here, this adjunct word verification unit 14 includes a means for testing the full possibility of the remainder of the kana character string as an enumeration of adjunct words, and a means for testing the remaining character strings as an enumeration of adjunct words. corresponds to means for outputting a kana adjunct character string separately from an independent word stem.

Furthermore, the independent word development section 11 includes a Japanese language dictionary index section 12.
Then, the Japanese language dictionary 13 is indexed to check whether an independent word exists with this attached word. At the same time, the adjunct string is tested for all possible combinations. If the two match and there is no shortage, the grammar analysis unit 1
0 takes independent word candidates one after another from the independent word development unit 11, tests the combination of the part of speech with the part of speech that is the first accessory word of the specified accessory word string, and if it is possible to combine the part of speech with the part of speech that is the first accessory word, The word character strings are combined and displayed on the display section 4 by the candidate display section 16. The operator uses the kana-kanji candidate selection section 3 to display the previous/next candidates and selects a matching kana-kanji character string. Further, 15 is an adjunct word dictionary searched by the adjunct word verification unit 14, 17 is a character font generator that generates character fonts, 6 is a text storage unit that temporarily stores sentences, and 7 is: This is a document storage unit that stores documents.

FIG. 2 is a schematic block diagram showing the hardware configuration of the present invention. The main CPU 51 performs calculations and
It controls the entire system. Further, the main CPU 51 has two slave CPUs, that is, a slave CPU 55 and a slave CPU 60.
Commands and data are exchanged between the main CPU 51, slave CPU 55, and slave CPU 60. The slave CPU 55 functions as an LCD controller that controls an LCD (not shown) for display, and further functions as a main CPU 5.
The contents of the ROM capsule 56 are read according to the instructions in step 1.
Load it into RAM53. On the other hand, slave CPU6
0 is a keyboard (not shown) and clock function,
Furthermore, it controls the power supply. Next, I/O port 54
connects with other peripheral devices, and an interrupt controller 59 controls interrupts from outside. Furthermore, 57 is a serial controller, and 58 is a timer. and,
The ROM 52 stores a program for starting the program, and a ROM capsule 56
When the ROM is stored in the ROM 52, the contents of the stored ROM are stored in
The program is loaded into the RAM 53 and execution of the program begins. The RAM 53 is used as a work area for the basic program stored in the ROM 52 or as a user area.

Further, the program for realizing the adjunct word test conversion of the present invention is a ROM stored in the ROM capsule 56.
is stored in

Next, the operation of the present invention will be explained using the flowchart shown in FIG. In the following explanation, the "level" of an adjunct word indicates the position of the adjunct word being processed in the kana character string. When the level of an attached word is 1, it means that the attached word is located at the beginning of the remaining kana character string obtained by cutting out independent words from the kana character string. Further, the adjunct word at the previous level means the adjunct word located before in the kana character string, and the adjunct word at the next level means the adjunct word that follows in the same way. Also, level 0 means a word located before the attached word of level 1, and actually means an independent word. Now, as an example, "Kyokara dake dake"
Consider the case of converting the sentence. First, the word "kiyo" is determined to be an independent word, and the sentence is divided as shown in FIG. 4a (<1>).
Next, the attached word is searched from the beginning in the character string "Karadakeda" and the longest matching word is selected (<2>). In this case, the longest adjunct is "kara"
Therefore, the character string is divided as shown in FIG. 4b. After <2>, the presence or absence of an auxiliary word is checked (<3>), and in this case, a part-of-speech combination test with "Today", which is a word at the next previous level, is performed (<4 >, <5>). In this case, since it is possible to combine, it is possible to check whether there are any remaining digits in the string (<6>). In this case, there are still digits left, so the next level adjunct "da" is extracted (<7>), and then the process returns to <2> and repeats the above process, <2>, <3>, <4>,
The process progresses in steps <5> and <6>, and since there are no more digits left, the adjunct word verification is OK (<8>), and the process ends.

Further, the process <10> is a process in which when an adjunct word is found but its combination with a word at the previous level is not successful, the digit of the adjunct word is shortened by one and the search for the adjunct word is started again.

In addition, if there is no matching result after searching for an adjunct word, first test whether the adjunct word is level 1 (<11>), and if YES, an adjunct word verification error (<12>) is detected. >), and the process ends. In this case, there is no attached word that connects the extracted independent word and the remaining character string. Also <11>
In this process, if the level of the attached word is not 1, the attached word of the previous level that has been determined is destroyed, its digit is shortened (<13>), and the search for the attached word is started again.

(Effects) As explained above, according to the present invention, by assuming that the independent word stem is a certain range of kana character strings,
After extracting the independent word stem, the remaining character strings are checked for adjuncts, which reduces the number of character strings to be examined, resulting in the following effects.

The adjunct word development area can be reduced.

The time required for adjunct language testing can be reduced.

The time required to expand attached words can be reduced.

Therefore, it can be said that the present invention saves storage area and development time for adjunct word processing in kana-kanji conversion. In the present invention, the means for testing the possibility of using the remainder of a kana character string as a list of adjuncts is such that even if it becomes impossible to combine adjuncts, the digit of the extracted adjunct is shortened by one and the unit is reused. , search for adjuncts, and even if there is no adjunct word, if it is not in the first position, the previous adjunct word that was determined will be destroyed and the position will be returned to that digit. shorten and search for attached words again. Therefore,
According to the present invention, it is possible to precisely extract attached words from a kana character string, and it is possible to significantly improve the efficiency of kana-kanji conversion processing.

[Brief explanation of the drawing]

FIG. 1 is a functional block diagram schematically showing the kana-kanji conversion device of the present invention. FIG. 2 is a diagram showing the hardware configuration of the present invention. FIG. 3 is a processing flow chart of the apparatus of the present invention. FIG. 4 is a diagram for explaining the verification of attached words according to the present invention. 1...CPU, 8...Kana-Kanji conversion device, 9...
...Kana character string input section, 10... Grammar analysis section, 14
...Adjunct language test department.

Claims (1)

[Scope of Claims] 1. In a kana-kanji conversion device that extracts independent words and attached words from a kana character string input by key input and performs kana-kanji conversion, specifying an independent word length for the kana character string; A means of assuming that the independent word stem from which an independent word is extracted is a certain range of kana character strings, and the following processing (a), (b), (c), (d), and (e), i.e., (b) extracting the above from the kana character string. Search for attached words from the beginning of the remaining character string after cutting out the independent words, cut out the longest matching word as an attached word, and if there is an attached word, process (b).
If not, perform the process in (E). (B) Perform a part-of-speech combination test between the attached word extracted in (B) and the preceding word, and if the combination is possible, (d) Perform the following process, and if the combination is not possible. Perform the process of (c), (c) shorten the digit of the extracted attached word by one.
Return to (B) and search for adjunct words. (d) Check whether there are any remaining digits in the above kana character string and if there are no remaining digits, complete the adjunct word verification.
Finish the process, and if there are any remaining digits, return to (B) to extract the following attached word. (E) Verify whether the attached word is at the beginning, and
If it is the first adjunct word, it will be treated as an adjunct word verification error and the process will end. If it is not the first adjunct word, the previous adjunct word that was determined will be destroyed and its digit will be shortened.
Go back to (b) and search for adjuncts, and use the remainder of the kana character string as an enumeration of adjuncts to test its possibility; Test the part of speech of the kana adjunct character string that has been verified to be an enumeration of adjuncts by means of testing its possibility as an enumeration of adjuncts, and if it is possible to combine, use the adjunct string. 1. A kana-kanji conversion device comprising means for combining and displaying on a display unit.