JPS60195664A - Kana-to-kanji-converting system - Google Patents

Abstract

PURPOSE:To prevent separation errors in the separating operation of Kanji (Chinese character) from Kana (Japanese syllabary), by inputting separating information together with a solid-inputted sentence and performing Kana-to-Kanji conversion in accordance with the separating information. CONSTITUTION:A dictionary memory contains dictionaries necessary for making clause separating process, selecting process for selecting words corresponding to the reading sound of clauses, Kana-to-Kanji converting process, etc. At the time of the process, a Kana character string to be converted is inputted in a buffer memory 3 together with a separating mark in the unit of one sentence by using the Kana Key and separating mark slashing key of an input section 4 while watching the display of a display section 5. Then the conversion key of the input section 4 is operated and existence of the separating mark is discriminated. When the separating mark exists, the inputted Kana sentence is divided into fractions and Kana-to-Kanji conversion is then performed on the solid-inputted Kana sentence. After clause separating process is performed in the unit of clause separated by the separating meark by referring to the dictionary and table in the dictionary memory 2 necessary for making the separating process the Kana-to-Kanji conversion is performed in the unit of clause.

Description

[Detailed Description of the Invention] Technical Field of the Invention The present invention relates to an improvement of a kana-kanji conversion method, and more particularly, the present invention relates to an improvement of a kana-kanji conversion method, and more particularly, to a kana-kanji conversion device for converting a plain input sentence of kana characters into a sentence containing kanji. This invention relates to a kana-kanji conversion method that improves conversion efficiency.

Technical Background of the Invention and its Problems> Recently, there has been a system that automatically converts sentences containing kana and kanji by inputting the pronunciations including punctuation marks, without requiring the input of additional information such as phrase delimiter marks. A kana-to-kanji conversion device has been put into practical use that outputs the kana-to-kanji characters (hereinafter referred to as the solid input method or solid input kana-to-kanji conversion method).

However, in this type of solid input kana-kanji conversion method, from a solid input sentence of kana characters without a break mark such as a 5-bunse separation mark, the device performs segmentation processing to determine the location of the segment. Since the kana-kanji conversion process is performed based on the determined break position, an error may occur in determining the break mark insertion position, which may result in erroneous conversion.

For example, [It's not spring yet in the northern countries. ” I expected the output to be “Kita uni hamada haruni hanatteinai.

If you enter the kana string ``'' and perform the conversion process, ``
The northern country has not released it in spring yet. '', but such erroneous conversions cannot occur in the segmentation method of kana-kanji conversion, and are unique to the solid input kana-kanji conversion method.

<Object of the Invention> The present invention has been made with the aim of eliminating the above-mentioned conventional problems and improving the kana-kanji conversion rate.In order to achieve this purpose, the kana-kanji conversion method of the present invention is An input means for inputting a Japanese sentence as a solid input sentence of a kana sentence and inputting division information, and a unit of kana strings separated by the input division information of the kana sentence input by this input means. It involves a separation process that converts it into a sentence containing kanji. The apparatus is equipped with an input kana-kanji conversion means, so that by inputting division information together with a solid input sentence, division processing is executed based on the input division information to prevent division errors. ing.

<Embodiment of the Invention> Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a functional block diagram showing the configuration of a kana-kanji conversion device implementing the kana-kanji conversion method of the present invention.

In FIG. 1, 1 is a control unit, 2#-i dictionary memory unit,
3-digit buffer memory section, 4-digit keyboard input section, 5
The control unit 1 is constituted by, for example, a microcomputer, and the other components 2 to
5, and also performs various processes described below using the components 2 to 5. The dictionary memory unit 2 includes dictionaries necessary for segmentation processing, homophone selection processing, kana-kanji conversion processing, etc.

FIG. 2 Vi shows a kana sequence and positions for inserting parting marks (■ to ■) when converting the sentence ``Hokkaido tri, it's not yet spring.'' by solid input.

For example, if a solid input conversion is performed without inserting any one-minute mark, the following erroneous conversion may occur.

``The northern country has not been released/in spring yet.'' Note that the ``/'' in the above sentence indicates the position where it has been automatically divided.

For this, instead of inputting a solid kana string, according to the present invention, if you input a kana string with a dividing mark added at the position of ■ in Figure 2 and perform the conversion process, you will get "Hokugoku i11/l/ It's not spring yet.''
Correct conversion is achieved, and there is no need to specify divisions at four locations (① to ①) as in the conventional split input method, and the number of operations is reduced, allowing correct conversion results to be obtained.

In Figure 2, there are four positions where you can specify divisions, but it is generally difficult to determine the positions where you can specify divisions, as shown in Figure 3. This is also the case when a continuous part or a character at the boundary of a clause is absorbed into a subsequent clause as in the above example, or vice versa as shown in FIG.

In conversion using solid input, such division failures may occur, but there is a possibility that position determination may fail within 1 minute. Efficient conversion is possible by adding separation marks according to the present invention.

In particular, in Figure 4, the output of b) is also correct as a sentence,
Since only the operator can determine whether a) or b) are correct or incorrect, the presence or absence of a separation mark has an important meaning.

In addition, r, "J" and "3" in FIGS. 3 and 4
The "■" mark in the figure is drawn for convenience of explanation, and "-r/" indicates the phrase boundary after kana-kanji conversion.
J indicates that it is a suffix. Also the third
A) and b) in the figure and FIG. 4 respectively show the conversion results that can be expected when a) is partially added with a dividing mark according to the present invention, and b) is without a dividing mark. An example of the conversion pot effect in case of flat input is shown.

Further, the inputting process for dividing marks according to the present invention may also be used as a dividing mark entry key.

Next, the processing operation of the apparatus implementing the present invention will be explained according to the processing flow shown in FIG.

First, in step n1 of FIG. 5, the kana character string to be converted, for example, "Harunihanatte..." is displayed in the buffer memory 3 on the display section 5 using the kana key and the dividing mark key of the input section 4, one sentence at a time, together with a dividing mark. Enter while looking at the display. At this time, the input kana character string is displayed on the display section 5 together with a dividing mark.

To enter the dividing mark, use "Space" or "Sura Nosh"
Symbols that are not often used in actual input/output may be used, such as "slash" as an example, and t/i in Figure 2 as the input position of the dividing mark.
If you input it to the position, the screen of the display unit 5 will show [Haruniha/Nattenai]. ] The input characters are displayed as shown below, and the input process for one sentence (step nl) is completed.

Next, in response to a keystroke instruction such as a "conversion key" from the input unit 4, the process moves to step n2.

In step n2, it is determined whether the operator has input a separation mark, and if there is a separation mark, step n
3, the input kana sentence is subdivided using dividing marks, and the process proceeds to step n4, where the input kana-to-kanji conversion processing is performed. After segment segmentation processing is performed while referring to dictionaries, tables, etc. necessary for processing, kana-kanji conversion processing is performed for each clause. Note that the bunsetsu segmentation process and the kana-kanji conversion process themselves are conventionally known processing techniques, so detailed explanations thereof will be omitted.

In the conventional solid input, there is no input of a half-minute mark and there is no subdivision processing of the kana sentence using the break marks, and the process directly moves from step nl to step n4.

In step n4, the unsubdivided or subdivided kana string is sequentially converted, and in step n5 it is determined whether or not the subdivided sentence remains, and the input kana string is Conversion processing continues until there are no more. Here, although the kana sentence is subdivided, it does not need to be divided into clauses at all, and even if the kana sentence converted to the rope culture is longer than two clauses, it is natural that the kana-kanji characters input in step n4 Processing is done by conversion.

It is also clear that the conversion process is performed in step n4 even if there is no separation mark at all.

When all of the input kana strings have been converted, this state is determined in step n5, and the process moves to step n6 to display the results. The operator looks at this result and decides whether to make a correction or input (step n7). If necessary, go through the correction process in step n8, and if not necessary, directly return to LOOP 1 and repeat the entire process. It turns out.

Furthermore, if the correction process is not specified in step n7, the process returns to step n1 via LOOP 1 by inputting a kana, so the input of two sentences may be repeated, and correction work may be performed later.

In Fig. 6, input h in the operation of 11 steps n1.
This shows the contents of the buffer that stores the input kana string with a dividing mark added.Step n3 is performed so that the conversion result always reaches the clause boundary at the position of the dividing mark "/" in the control unit 1-digit buffer 11. , n4, conversion result candidates as shown in, for example, ]2 to 14 are obtained in the conversion results P and N4. The most suitable candidate is selected from the obtained candidate group in the process of converting solid input kana to kanji in step n4. As for this selection method, for example, in the case of the contents of conversion result buffer 12, one Kanji character and one bunsetsu are considered inappropriate, and in the case of conversion result buffers 2 and 14, candidates with a large number of bunsetsu are considered inappropriate if they are the same kana string. The contents of the buffer 13 can be selected as the conversion result using an appropriate selection method.

Figure 7 is by using the dividing mark.

FIG. 3 is a diagram showing the storage contents of a temporary buffer for explaining another embodiment of the kana-kanji conversion method of the present invention that prevents erroneous conversion, and is a partial extraction of the kana string in FIG. 2;

Fig. 7 shows the contents of the buffer memory of a kana string input in 21 digits, and shows the contents of the conversion result buffer when the kana string in buffer 21 is converted in 22 and 23fi respectively with solid input. .

During the analysis of the solid input, two types of analysis results, such as the contents of the conversion result buffers 22 and 23, are extracted. Furthermore, the contents of the buffer 21 for the kana string hold the correspondence shown by the broken line. Furthermore, conversion result buffer 2
2 and 23 have information on phrase breaks based on solid input kana-kanji conversion, and these phrase breaks are conceptually shown as E].

In FIG. 7, 24 indicates the content of the buffer of the kana string when the dividing mark "/" is added according to the present invention, and the buffers 22 and 23 of the conversion result are "reading".
In terms of "notation" and "notation", the correspondence is the same as when converting the kana string in the buffer 121.

The kana string in the buffer 24 is converted into the conversion result buffers 22 and 23.
The conversion result buffer of 23 is selected by comparing the contents with the content of the conversion result of 23 by fully verifying whether the break mark "/" added to the kana string of the buffer 24 matches the bunsetsu boundary of the conversion result buffer y22-23. , display it as output.

<Effects of the Invention> As described above, according to the present invention, in addition to solid input sentences, segmentation and
Since the system is structured so that the Kana-to-Kanji conversion is performed based on the dividing information, the Kana-to-Kanji conversion technology for solid input is used, and if there is a mark that tends to be divided, the dividing mark is used. You can prevent sharing mistakes by doing this. Furthermore, according to the present invention, since there is no need to input segmentation information for each clause, it is possible to absorb the fluctuations that tend to occur in clauses, and to effectively utilize the features of the solid input method for inputting sentences with fewer key operations. I can do it.

In addition, operators who are accustomed to the conventional bunsetsu segment input method can also use the conventional input method.

It is clear that the separation mark is emphasized and converted, and the present invention incorporates the input method from the bunsetsu segment.

Furthermore, it is clear that punctuation marks, symbols, changed positions of character types (katakana, alphanumeric characters), etc. that are unconsciously inputted by the operator due to a request for outputting a sentence can be used as one-minute marks. .

[Brief explanation of drawings]

FIG. 1 is a functional block diagram showing the configuration of a kana-kanji conversion device implementing the kana-kanji conversion method of the present invention, FIG. 2 is a diagram showing an example of an input kana string, and FIGS. 3 and 4 are respectively A diagram showing a display example of a conversion result and the positional relationship with a dividing mark, FIG. 5 is an operation flow chart for explaining the processing operation of an embodiment of the kana-kanji conversion method of the present invention, and FIG. A diagram showing the memory contents of the buffer 1 that provides the operation theory EfAK,
FIG. 7 is a diagram showing the stored contents of a buffer for explaining the operation of another embodiment of the present invention. 1: Control unit, 2: Dictionary memory unit, 3: Buffer 1
Memory section, 4... input section, 5... display section. - Shizuichi°1・(Representative Attorney Patent Attorney Aihiko Fukushi (and 2 others) Figure 1 Figure 2 a) I Lk X': J'F, 71? l i' input, -' r expression 'n, /ell>v,' %6. Figure 3 Q) ``H,'t'M r< ,,l in,?,'b
$ia. b) TV,'/ltW+:+x,-'iT,r
,='lh strong Te. Figure 4 Figure 5111+ Figure 6 Figure 7

Claims (1)

[Claims] 1. An input means for inputting a Japanese sentence as a continuous input sentence of a kana sentence and inputting division information, and a kana string separated by the input division information of the kana sentence inputted by the input means. comprising a solid input kana-kanji conversion means that involves division processing to convert into a sentence containing kanji as a unit,
A kana-kanji conversion method characterized in that parting information is input together with a solid input sentence, and kana-kanji conversion is performed based on the parting information.