JPH04552A - Kana-kanji conversion device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a kana-kanji conversion device that converts a character string input in the form of kana into a string containing kana.

[Prior Art] In recent years, kana-kanji conversion devices have been widely used in various Japanese language processing devices, including Japanese word processors, as devices used to input Japanese characters. The kana-kanji converter is used to convert an input kana string into a kana-kanji mixed string. For example, if you input the following kana string ``Asu wa tenkigayo (naru)'', it can be converted to ``tomorrow the weather will be nice (naru)''.

[Problems to be Solved by the Invention] However, the conventional kana-kanji conversion device described above does not have a mechanism to incorporate into the kana-kanji conversion analysis linguistic phenomena that occur prominently in a particular field. There was a problem that conversion analysis could not be performed. For example, when writing a letter, if the user inputs the following kana string ``Sono Greetings'', it is grammatically correct as ``Greetings afterward (Sono Greetings)'', but when writing a letter, the user This will result in an unnatural conversion result that is contrary to your intention. This unnaturalness is caused by the fact that when writing letters, there is no way to reflect in the conversion result the fine-grained kana information for each field, in which "go~ (suru)" is especially used as a prefix of respect. Therefore, in the above example, it depends on the previous learning, dictionary order, evaluation value of the phrase pattern connection table, etc., and "Greetings afterwards" is selected as the conversion result.

[Means for Solving the Problem] In order to solve this problem, the present invention provides a phrase pattern evaluation table that stores data indicating the grammatical certainty of connections between phrases for each field, and It is configured to determine the priority of kanji data corresponding to input kana data using a bunsetsu pattern evaluation table corresponding to the field.

1 Effect] With the above configuration, kanji data conforming to a phrase pattern suitable for a designated field is preferentially output.

[Embodiment] Hereinafter, a kana-kanji conversion device according to an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a functional block diagram of a kana-kanji conversion apparatus according to an embodiment of the present invention, and l is an input means having means for inputting a kana string and for instructing conversion from a kana string to a kanji-kana mixed string. Reference numeral 2 denotes an input kana string storage section that stores the kana string inputted from the input means 1. 3 is a word dictionary that stores word pronunciations, kanji notation, and part-of-speech information in pairs.

4 is a clause pattern evaluation table for each use mode, which indicates the degree to which each clause pattern as shown in FIGS. 5(a) and 5(b) is likely to be a clause. 5
is a phrase pattern classification section that converts the kana string stored in the input kana string storage section into a kanji-kana mixed string and classifies the phrase patterns obtained as analysis candidates. 6 searches the word dictionary 3 for the kana string stored in the input kana string storage section 2 and converts it into a string containing kanji and kana, and exchanges the clause candidates obtained as a result of the analysis with the clause pattern classification section 5. The kana classifier classifies the bunsetsu patterns, and then selects the kanji-kana-mixed string with the most probable bunsetsu pattern in the usage mode from among the multiple kanji-kana-mixed string candidates while referring to the bunsetsu pattern evaluation table 4, and outputs its notation. This is a kanji conversion control unit. Reference numeral 7 denotes an output character string storage unit that stores the kanji-kana mixed string output from the kana-kanji conversion control unit. Reference numeral 8 denotes a display means for displaying the kanji-kana-mixed string stored in the output character string storage unit as the conversion result.

FIG. 2 is a circuit block diagram of the kana-kanji conversion device of this embodiment, where 9 is a keyboard, 10 is a central processing unit (hereinafter referred to as
11 is a cathode ray tube display (hereinafter referred to as CRT display), 12 is a read-only memory (hereinafter referred to as ROM), and 13 is a random access memory (hereinafter referred to as RAM). The input means 1 shown in FIG. 1 is provided by a keyboard 9, the input kana string storage section 2 and the output character string storage section 7 are provided by a RAM 13, and the word dictionary 3, bunsetsu pattern classification section 5, and bunsetsu pattern evaluation table are provided by a ROM 12. The kana-kanji conversion control unit 6 is realized by the CPU 10 exchanging data with the ROM 12 and RAM 13 while executing a program stored in the ROM 13, and the output means is realized by the CRT display 10.

The operation of the kana-kanji conversion apparatus of this embodiment configured as described above will be explained below based on the flowchart shown in FIG.

First, in step (a), when a kana string is input from the input means, the input section sends the contents to the kana-kanji conversion control section and instructs conversion into a sentence containing kanji and kana. For example, when the next kana string "Greetings" is input from the input means, the above kana string is stored in the input kana string storage section 2. Next, when the conversion key is input from the input means 1, the process proceeds to step (b), where the kana-kanji conversion control unit 6 receives the conversion instruction from the input means 1, searches the word dictionary 3, and converts the phrases and phrases corresponding to the input. We will decide the boundaries between. In the case of the example sentence, the kana-kanji conversion control unit 6 first breaks down the input sentence into clause units: (1) "Sonogo/Greetings J" (2) "Sono/Go/Greetings". Furthermore, in step (c), the kana-kanji conversion control unit 6 writes candidates for each clause that has just been decomposed into the output character string storage unit 7. The contents of the output character string storage section 7 at step (c) are shown in FIG. 4(a). Subsequently, in step (d), the phrase pattern classification section 5 refers to the output character string storage section 7, classifies the phrase patterns, and writes them into the output character string storage section 7. In the case of this example sentence, pattern (1) (: adverb "after" ten words alone "greeting")
It is classified as a bunsetsu pattern of pattern (2) (-adnominal ``sono'', prefix ``go'', ``ju-respect'', ``greeting'' alone). The contents of the output character string storage section 7 at the stage of step (d) are shown in FIG. 4(b). In step <E), a clause pattern evaluation table of the specified usage mode is prepared among the plural clause pattern evaluation tables. In this example, this is the evaluation table for letter text mode. Figure 5 (
Figure a) shows the contents of the clause pattern evaluation table in letter mode. Note that unless the user specifies otherwise, the normal evaluation table 5(b) shown in FIG. 5(b) is used as the evaluation table. In step (v), it is determined whether evaluations have been determined for all clauses, that is, whether all evaluation values in the output character string storage section 7 have been given. If the evaluation values of all clauses have been determined, proceed to step (S); otherwise, proceed to step (G). In step (g), the kana-kanji conversion control unit 6 determines the evaluation value of the evaluation clause by referring to the evaluation table as shown in FIG. An evaluation value is given to the conversion candidate phrase in section 7. In this example, if the phrase pattern is ``respected prefix master's formal expression alone'', the evaluation table for letter mode is referred to and the evaluation value 4 of the phrase pattern ``respected prefix master formal expression alone'' is output as a character string. Write it as the evaluation value of [Phrase No. 2, notation "Greetings"] in the storage section. From then on, repeat steps (to) to (g) until all clauses are given evaluation values.
Repeat. Figure 4 (
Shown in c). In step (ch), the kana-kanji conversion control unit 6
is the output kana string storage unit 7 (Phrase pattern evaluation table 4)
The phrase with the phrase pattern with the highest evaluation value is found by referring to , and the output means 8 outputs it as a notation for the hiragana string.
Output to. In other words, in this example, the evaluation value for the conversion candidate kanji/kana mixed string "Greetings afterwards" is the phrase "Afterwards J"
"Greetings" is determined by the sum of the evaluation values for each. Similarly, an evaluation value is given to the conversion candidate "Sono-Greetings", and "Sono-Greetings" with the highest evaluation value as a conversion result for the hiragana string "Sono-Greetings" is the best one for the current usage mode, that is, letter mode. It will be selected as a result of conversion.

[Effect of the Invention J As described above, the present invention uses a clause pattern evaluation table for each use mode, which stores data indicating the grammatical certainty of the connection between bunsetsu and bunsetsu for each field. Since the priority of the kanji data corresponding to the input data is determined, it is possible to provide a kana-kanji conversion device with excellent conversion efficiency.

[Brief explanation of drawings]

FIG. 1 is a functional block diagram of a kana-kanji conversion device according to an embodiment of the invention, FIG. 2 is a circuit block diagram of the same device, FIG. 3 is a flowchart showing the control procedure of this embodiment, and FIG.
5A, 5B, and 5C are diagrams each showing the contents of the output character string storage section 7, and FIGS. 5A and 5B are diagrams each showing a clause pattern connection table. 1... Input means, 2... Input kana string storage section, 3...
・Word dictionary, 4...Bunsetsu pattern evaluation table, 5・
... Clause pattern classification unit, 6... Kana-kanji conversion control unit, 7... Output kana string storage unit, 8... Name of output means agent Patent attorney Shigetaka Awano and one other person Figure 1 Paper ℃ 1゛〉

Claims (1)

[Scope of Claims] An input means for inputting kana data, a storage means for storing the kana data inputted from the input means, and kanji data and grammar data corresponding to the kana data constituting a word as headings. a dictionary, a clause pattern evaluation table storing data indicating the grammatical certainty of connections between clauses for each field; a specification means for specifying a domain; A kana-kanji conversion device comprising means for determining the priority of kanji data corresponding to kana data stored in the storage means using a phrase pattern evaluation table.