JPS5887618A - Kana-kanji conversion processing method - Google Patents

Abstract

PURPOSE:To reduce the retrieval time up to the 1st stage and to improve the processing speed of document forming work, by providing two stages of a converted candidate retrieval range, extending the range at the 2nd stage if characters that the operator does not intend to output are displayed one after another at the 1st stage, and performing the conversion again. CONSTITUTION:An input preprocessing section 1 recognizes alphanumericals and clause delimiting information in inputted Kana sentences to extract Kana character strings and gives the Kana character strings to a word extracting section 2 as a conversion unit under controls of a conversion controlling section 5. The sections 2, 5 and a language controlling section 3 refer to a disctionary to extract the candidates of word strings permitted as the grammer and sentences mixed with Kanji and Kana are outputted from an output controlling section 4. Inextracting up the word strings, the conversion candidate retrieval range is defined as two states, and if characters that the operator does not want to output are displayed one after another at the 1st stage, the range is extended at the 2nd stage and the candidates are extracted through the conversion again. The retrieval time up to the 1st stage can be reduced and the processing speed of the document forming work can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a kanji conversion processing method, and particularly relates to a kana-kanji/banji conversion processing method that can improve the processing speed of a long calligraphy work in a kana-kanji conversion processing method that provides phrase break information.

Various methods have been proposed in the past for processing methods for converting Japanese characters into Japanese characters (hereinafter simply referred to as "conversion"). Although there are differences in conventional conversion methods in terms of text, clause designation, word designation, etc., in general, all possible conversion character strings for an input kana character string are searched and the conversion results are output and displayed. and homophones (hereinafter referred to as "homonyms")
If there is a conversion string that is not intended by the operator, it displays the conversion string from the most likely candidates and asks the operator to select and confirm. By pressing the first candidate key, the excerpt and another converted character string are alternately displayed. This P! ! If there is something the operator intended at 1m, the operator presses the selection key and the dyed conversion character string is determined.

However, the above-mentioned method always searches for all possible conversion strings, so it has the serious drawback that it takes a long time to display conversion candidates for the first time. .

The present invention has been made in consideration of the above-mentioned circumstances, and its purpose is to eliminate the above-mentioned drawbacks of conventional conversion methods, shorten the time required to display the first conversion, and improve user-friendliness for operators. The purpose of the present invention is to provide a kana-kanji conversion processing method that improves handwriting and improves the processing speed of document creation work.

The purpose of the present invention is to provide a kanji conversion processing method in which an input sentence is broken down into division units centered on independent words, input as kana characters, and corresponding sentences containing kanji and kana are sequentially obtained. By using the two-stage conversion candidate search range, if the converted character string output and displayed as the most likely conversion candidate in the first conversion candidate search range is a character string that the operator did not intend, the runner-up The first next candidate key outputs and displays the long-sleeved kimono in the next layer, and if the character string that the operator intends is not found among the second candidate keys, the second next candidate key is used to display the conversion candidate search range. This is achieved by a kana-kanji conversion processing method characterized by expanding the search range to the next conversion candidate search range and performing the conversion operation again.

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a diagram showing kana-kanji conversion processing according to an embodiment of the present invention. In the figure, 1 is an input preprocessing unit, 2
is a word extraction section, 3 is a homophone candy discrimination section, number is an output control section, and heart is a conversion control section.

When a Japanese sentence is input as a kana sentence, it is output as a mixed kanji/kana sentence through the following processing.

The human preprocessing unit 1 separates alphanumeric characters and phrases in the input kana sentence)
The information section is processed by ¥gIII to extract the kana character string that is subject to political conversion, and is sent to the word extraction section 2 under the control of the conversion control section 6.
Waves the kana character string that is the conversion unit.

The word extraction unit 2ti extracts the kana characters from the kana character string by referring to a word dictionary with kana as heading, a dictionary containing part of speech information of the part of speech added to the word dictionary, connection information for each crystal body, etc. Attempts to match columns and dictionary entries, and extracts grammatically acceptable single column candidates. Homophone candy discrimination part 3
When there are multiple candidates for the word string, the most likely single candy i is determined using frequency information of the words.
Output 1lIJ141 section 4Fi displays the corresponding characters of the phrase determined as described above on the output display device. Thereafter, when the manual preprocessing section 1 receives a familiar number designating the next candidate for the master control key, the runner-up candidate is displayed instead of the most likely iron assistant that was initially displayed.

Word extraction bli2, which is the main point of the present invention, will be explained below.

Single candy extraction unit F'1 Usually extracts single strings based on independent word matching using the so-called longest match method. Once, by Lt combination of a certain major 6 independent song, among the converted fusode group extracted from that clause, its independent i
A conversion string whose constituent element is an independent word shorter than u cannot be given. As mentioned above, the maximum method has a downside in that it does not necessarily extract all possible cases, but it narrows down the TFA conversion arm ice detection yoke by passing through another destination.
There is also the positive side of being able to reduce the number of k-replacement candidates.
In addition, this method can yield results for most of the poor exchanges in Sk, and the review time for sister plans can be shortened, so it is the best choice for the first search for barbaric exchange candidates ffi f. It's a matter of choice.

Therefore, in the kana-kanji conversion process on the side of this solid line, the first
By pressing the III key input to the input processing unit shown in the figure, a conversion candidate search kifl update key (hereinafter referred to as "z1 category key") is added to the next candidate key that specifies the next candidate and company regulations.
If there is no M-complement desired by the operator among the candidates obtained by searching within the so-called first conversion candidate search range using the longest match method, press the new range key all the way; Second modified candidate search range H&: The search range is changed and further search is performed.

In this case, a word extraction method called a so-called brute force method is used as the second conversion candidate search range.

With this extraction method, even if matching is successful during the search process,
This method continues the search and extracts candidates that all have "T n'Q properties." Therefore, if the candidate power desired by the operator is found at the beginning of the search, the operator will perform an extra search and perform V after that, until the first conversion candidate is displayed. It takes time to do this, but if you combine this with the -% longest - teaching method,
This is a very effective method when used as a method to search from one point of the blow when the operator cannot obtain the candidate desired by the search using the I&L teaching method.

In addition, in the conversion process of the above embodiment, providing a new range key separately from the next candidate key has the following advantages. In other words, normally when you press the next candidate key, new conversion candidates are displayed one after another, but if you press the next candidate key again when all the conversion candidates have been found, the first conversion candidate is displayed again. It looks like this. This is convenient in that even if the operator makes a mistake in selection at this stage, all conversion candidates can be immediately reviewed. This makes it possible to proceed to the search in the second conversion candidate search range even after confirming that the operator does not have the desired one among the searched candidates.

In the above embodiment, the first and second conversion candidate search ranges are an example of using the &1 search using the longest-teaching method and the search using the brute force method, respectively, but this is not limited to this. It goes without saying that the search range switching method #'i may also be used.

As mentioned above, according to Honbatsu Mei, an input sentence is broken down into separate units centered on independent words, input as kana characters, and the corresponding sentences containing kanji and kana are sequentially obtained using a kana-kanji conversion processing method. , a conversion candidate search range of at least two stages is provided, and when the converted character string output and displayed as the most likely candidate in the first conversion candidate search range is a character string not intended by the operator, the runner-up candidate is determined. The first
The next candidate key is used to output and display one after another, and if the character string intended by the operator is not among the runner-up candidates, the second candidate key is used to convert the conversion candidate sleeve search range to the next candidate. Since the conversion operation is performed by expanding the candidate search range, the time until the first conversion candidate is displayed is shortened, improving usability for the operator and speeding up the processing of document creation tasks.

[Brief explanation of drawings]

Figure 1 is a plotter diagram of kana-kanji conversion processing, which is an embodiment of the present invention. 1: Input preprocessing section, 2: Word extraction section, 3: Homophone discrimination section, 4: Output control section, ro: Conversion control section. Patent applicant: Co., Ltd. 1) 25. . Agent Masaru Isomura −・ −1, ″l

Claims (1)

[Claims] In the kana-kanji conversion processing method, an input sentence is broken down into division units centered on independent words and input in kana characters, and corresponding sentences containing kanji and kana are sequentially obtained. At least two stages of conversion candidate search range are established. If the converted character string output and displayed as the most likely conversion candidate in the first conversion candidate search range is a character string not intended by the operator, the next candidates are output one after another using the first next candidate key. In addition, if the character string intended by the operator is not among the runner-up candidates, a second candidate key is displayed.
A kana-kanji conversion processing method characterized in that the conversion candidate search range is expanded to the next conversion candidate search range by - and the conversion sweep is performed again.