JPS62214468A - Kana-kanji converter - Google Patents

Abstract

PURPOSE:To quickly perform the whole of the Kana (Japanese syllabary)-Kanji(Chinese character) conversion processing by using a dictionary for grammatical processing constituted from a word dictionary in an external storage device to write clauses divisionally at a high speed and using this word dictionary only for conversion to Kanji character strings. CONSTITUTION:When a Kana-Kanji converter 7 is started, a dictionary 11 for grammatical processing is generated from a word dictionary 12 by a grammatical processing dictionary generating part 8. When the character string read in from an input part 2 is sent to the Kana-Kanji converter 7, a clause divisional writing part 10 retrieves the dictionary 11 for grammatical processing to write clauses divisionally. Candidates of clause divisional write data are sent to a Kanji candidate generating part 13, and clause candidates are generated while retrieving the word dictionary 12, and candidates are outputted successively in clause units by a clause candidate output part 14. The clause candidate group subjected to Kana-Kanji conversion is displayed on a clause candidate display part 4, and clauses are selected and settled by a clause candidate selecting part 3 and a clause canditate settling part 5 and the stored in a sentence storage part 6.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a word dictionary search in a kana-kanji conversion device that converts an input kana character string into a kana-kanji mixed sentence in order to input a Japanese sentence.

[Conventional technology]

Conventional methods for speeding up word dictionary searches in kana-kanji conversion have been roughly divided into the following methods.

(1) Build a word dictionary on a high-speed retrieval medium (ROM).

(2) As many search tables (indexes) for word dictionary searches as possible reside in the main memory.

(6) Create a tree structure for the word dictionary to reduce the number of accesses.

However, with method (1), complete word learning based on frequency of use is impossible due to the read-only storage device, and (2)
, Method (a) is more effective than processing Setsubun Kamichisho! Since fe is directly searched, there is a drawback that it takes a long time to search when a slow storage medium is used.

[Problem that the invention seeks to solve]

As described above, if a word dictionary is directly searched rather than a bunsetsu segmentation process, if a slow storage medium is used, the search time will be longer and the kana-kanji conversion process time will be increased. In particular, in the case of conversion methods that require a large number of dictionary searches, such as the "two-clause longest match method" and the "minimum clause count method," it is necessary to reduce the search time.

[Means for solving problems]

In order to solve these problems, the present invention provides a kana-kanji conversion device that uses a word dictionary (hereinafter referred to as "word dictionary") in an external storage device that is only used for segment separation (hereinafter referred to as "grammar processing"). (abbreviated as "dictionary") is constructed on the main memory.

For this reason, there is a means for creating a grammar processing dictionary consisting only of kana headings and parts of speech (grammar processing dictionary creation means), a clause dividing means using this grammar processing dictionary, and a unit in which the clauses are divided. A kanji candidate creation method that converts kana character strings into kanji character strings, and once a word is selected,
It has means (usage frequency learning means) for learning the usage frequency in both the word dictionary and the grammar processing dictionary.

[Effect]

Originally, the information required for bunsetsu separation is only the kana heading of the word and the part of speech, which is grammatical information.The information needed to convert a kana string into a kanji string is the grammatical information of the words searched for bunsetsu separation. The word dictionary on the external storage device is used only when it is considered suitable for the purpose of writing, and only when segmentation is performed at high speed in the main memory and when converting kana character strings to kanji character strings. This greatly speeds up the entire Kana-Kanji conversion process.

Therefore, with these grammar processing dictionary creation means, clause separation means, kanji candidate creation means, and usage frequency learning means,
Speed up the entire Kana-Kanji conversion process.

〔Example〕

Hereinafter, details of the present invention will be explained based on illustrated embodiments.

FIG. 1 is a diagram showing a kana-kanji conversion device of the present invention. Reference numeral 1 in the figure is a central processing unit (hereinafter abbreviated as 0'PU) that controls the entire system based on data input from the input section 2. When the kana-kanji conversion device 7 is started up, the grammar processing dictionary creating section 8 creates a grammar processing dictionary 11 from the word dictionaries 111 and 2, and then begins the kana-kanji conversion process.

The character string read by the input unit 2 is sent to the kana-kanji conversion device 7, where it is converted into kana-kanji. In the kana-kanji conversion device 7, a phrase separation section 10 searches the grammar processing dictionary 11 and performs phrase separation. This clause separation data candidate is sent to the kanji candidate creation unit 13,
Phrase candidates are created while searching the word dictionary 12, and the phrase candidate output unit 14 sequentially outputs candidates for each phrase.The phrase candidate group that has been subjected to kana-kanji conversion is displayed by the phrase candidate display unit 4, and the phrase candidates are output as phrase candidates. Laundry department 3L-h-3!
``h sunchufukasei τ女倖jun骪凛中郎 [对1-1-) 1 is determined. When all the phrases are determined, the data is stored in the sentence storage section 6, and the kana-kanji conversion device 7 Based on this, the usage frequency learning unit 16 determines the usage of the word in the grammar processing dictionary 11 and word dictionary 1112 based on the analysis data accumulated in the kanji candidate creation unit 13. Perform frequency learning.

This series of operations can reduce the word search processing time of the bunsetsu segmentation process, which requires the maximum processing time in kana-kanji conversion, and enables high-speed kana-kanji conversion.

FIG. 2 is a detailed illustration of the present invention.

Here, at 200, a grammar processing dictionary is created before the kana-kanji conversion begins.

At step 201, a kana character string is input, and if this is an end code, end processing is performed at step 202. If it is not completed, the phrase separation is performed in step 203 while searching/referring to the grammar processing dictionary 11, and based on this data, the word dictionary 12 is searched/referenced in step 204, and the kanji candidate data for the phrase separation is created. If the candidate is selected and confirmed by the user for each phrase in 205 and a cancellation code is input, the process returns to 201. If not, the usage frequency learning of the word dictionary is performed in 207, and in 208 The frequency of use of the word dictionary is learned, and the series of operations is repeated by returning to the manual input 201.

FIG. 3 is a diagram showing the structure of the grammar processing dictionary 11 and word dictionary of the present invention. The word dictionary is divided into blocks 12A as shown in the figure, and consists of a plurality of blocks. Kanji section 12b
and an index section 12α indicating the block,
The index part 12α is the maximum heading 12C′f of the block.
T is listed.

In addition, the kanji section 12b has a kana heading 12d2 a part of speech 126,
It has Kanji 12/ as a roux code.

The grammar processing dictionary only needs to have the information necessary for segmentation, so the data section 11 is divided into blocks as shown in the figure.
b (1jA consists of a plurality of blocks) and an index section 116 indicating the block, and the index section 11α lists the maximum headings 11Ce of the blocks. In addition, the data section 11A includes a kana heading 11d, a part of speech 11
It has g as a roux code and has no kanji information.

〔Effect of the invention〕

The present invention has the following effects.

(In Kana-Kanji conversion, this shortens the processing time of the clause separation process, which takes the most processing time, and improves the processing speed of Kana-Kanji conversion.

(2) Since the grammar processing dictionary does not contain kanji data, it has a relatively small capacity and can be constructed in main memory.

(3) Applying this idea, a word dictionary is placed on the host computer, a grammar processing dictionary is placed on the terminal computer, and only the required amount of kanji data is transferred to the host computer.
Applications such as transfer and use from a computer can be achieved.

[Brief explanation of drawings]

Figure 1 shows the Kana-Kanji conversion device of the present invention! FIG. 2 is a block diagram showing FIG. FIG. 2 is a flowchart showing the operation of the kana-kanji conversion device of the present invention. FIG. 5 is a diagram showing the structure of the grammar processing dictionary and word dictionary of the kana-kanji conversion device of the present invention. Applicant Seiko Epson Corporation Figure 2 1a 12c Figure 3

Claims (1)

[Claims] A kana-kanji conversion device that converts an input kana character string by selecting homophones and converting sentences containing kanji and kana. , a grammar analysis dictionary that extracts only kana headings and parts of speech is built on the internal storage device, and the grammar analysis dictionary is used for the clause separation process, and the kana character string after the clause separation process is converted into a kanji character string. A kana-kanji conversion device characterized in that it refers to and uses a word dictionary stored in an external storage device only at times.