JP2951486B2 - Kanji conversion device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a kanji conversion device, for example,
The present invention relates to a kanji conversion device suitable for a document creation device such as a Japanese word processor.

[0002]

2. Description of the Related Art FIG. 4 shows an example of a conventional kanji conversion device for converting a kana character sentence into a kanji mixed sentence. When the operator operates the character input unit 31, a kana character string is formed. The kana character string is supplied to the phrase analysis unit 32.

[0003] The phrase analyzing unit 32 forms phrase information composed of independent words and adjunct words based on a kana character string. The phrase information is supplied to the search unit 33 and the conversion unit 35. The search unit 33 searches the dictionary unit 34 for a kanji having a heading that matches the phrase information. And
Kanji information representing the kanji is formed. This kanji information is supplied to the conversion unit 35.

The selection of kanji in the dictionary unit 34 is performed as follows. For example, the phrase "the phrase"
It can be disassembled into an independent part and an attached part like "bunsetsu" + "o". Therefore, when the word "bunsetsu" is used as a heading and searched in the dictionary unit 34, a kanji "bunsetsu" is obtained, and kanji information corresponding to the kanji is formed. The kanji conversion is performed as described above, but the timing of dictionary search differs depending on the dictionary structure and the kanji conversion algorithm.

[0005] The conversion unit 35 converts the kanji into kanji as candidates based on the kanji information and outputs the kanji.
As a result, a bunsetsu output of a kanji character is obtained.

In order to improve the efficiency of kanji conversion,
As shown in FIG. 4, additional data such as part-of-speech information is attached to candidate kanji stored in the dictionary unit 34 to make the connection between the independent unit and the additional unit more reliable, or "Secretary of State" with semantic information on connection with
Various methods have been used to prevent erroneous conversion of a word such as "political time".

[0007]

In the prior art as described above, it is assumed that the operator has input a correct kana character string through the character input section 31. Therefore, there is a problem that an erroneous conversion occurs if an incorrect kana character string is input due to a key operation error in the character input unit 31.

Therefore, if the input kana character string is incorrect, there is a problem that the input kana character string must be corrected in order to obtain a correct character string. In general, character strings are input using a keyboard, which imposes a burden on beginners who are not accustomed to key operations to correct input characters, and improvements have been desired. .

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to convert an incorrect character string into an intended kanji without correcting the character string. It is an object of the present invention to provide a kanji conversion device for obtaining.

[0010]

According to the present invention, there is provided a character input unit for inputting a character, and a phrase analysis unit for decomposing an input character string into phrases to form phrase analysis information. A heading which is displayed as a character string and is provided with matching evaluation information, a dictionary in which at least a kanji corresponding to the heading is set as a candidate, and search means for searching the dictionary based on the sentence analysis information. A match detecting means for detecting a match between the phrase analysis information and the dictionary heading for each character, and subtracting the match evaluation information of the mismatched character from the maximum value of the match evaluation information to obtain a match degree; When the degree is equal to or more than a predetermined level with respect to the maximum value of the matching evaluation information, a conversion means for converting the sentence analysis information into kanji is provided.

[0011]

According to the present invention, each character of a heading set in a dictionary is provided with matching evaluation information, and the maximum value of matching evaluation information in the heading is obtained. Determines whether the kana character string input and decomposed into clauses matches the kana character string of the headline set in the dictionary for each character, and if there is no match, it does not match from the maximum value of the matching evaluation information The character matching evaluation information is subtracted. When the matching degree at the stage where the match search is completed is equal to or more than a predetermined level with respect to the maximum value of the matching evaluation information,
It is determined that the entered and decomposed kana character string matches the kana character string of the heading, and a candidate kanji corresponding to the heading is selected.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to FIGS. In the configuration of FIG. 1, the kanji conversion device includes a character input unit 1 as a character input unit, a phrase analysis unit 2 as a phrase analysis unit, a search unit 3 as a search unit, and a dictionary unit 4 as a dictionary. And a match detecting unit 5 serving as a match detecting means.
And a conversion unit 6 as conversion means.

A keyboard is used as the character input unit 1. Kana character strings input from the keyboard are:
It is supplied to the phrase analysis unit 2.

In the phrase analyzing unit 2, the input kana character string is decomposed into phrases, and the phrase is decomposed into an independent part and an attached part to form phrase analysis information. The kana character string is disassembled into clauses and the clauses are disassembled into self-supporting parts and adjunct parts by known methods. This phrase analysis information is supplied to the search unit 3 and the match detection unit 5.

The search section 3 selects a heading for each phrase based on the matching evaluation information described later.

The match detecting section 5 performs the processing shown in the flowchart of FIG. 3, and selects a specific heading whose matching degree is equal to or higher than a predetermined level from the selected headings.
This information is supplied to the conversion unit 6.

In the conversion unit 6, a kanji is selected based on the information supplied from the coincidence detection unit 5. As a result, a bunsetsu output of kanji characters is obtained.

In this embodiment, as shown in FIG. 2, a weight (referred to as conformity evaluation information in this specification) is applied to each character of a heading stored in the dictionary unit 4. Is attached. This matching evaluation information is provided to perform an accurate Kanji conversion by an ambiguous match search even if a part of a kana character string is missing. The matching evaluation information is composed of two bits and represented by four values.

Hereinafter, the regularity of the matching evaluation information will be described. (1) The weight of the first few characters of the heading character is increased. The weight of the character at the head of the heading is increased in order to narrow down the dictionary search block for the heading. That is, if the keys are not narrowed down to some extent, the number of search blocks becomes enormous, and the dictionary search time exceeds the practical range.

(2) The prompting sound, the repetitive sound, and the long code have a low information weight. For example, a computer ……………………………………………………………………………… ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••] It is.

(3) If the number of characters in the heading is small,
The weight is increased for each character. For a heading of three characters or less, the weight is increased because each character is necessary in the vocabulary structure. For example, characters with a small number of headings, such as characters (Mori, peach, gate, person ...), electronic (electro, electromagnetic, battery), are kanji that are candidates. This is because there are many.

Hereinafter, the operation of the coincidence detecting section 5 will be described with reference to the flowchart of FIG. In step S11, the maximum value of the matching evaluation information is calculated. That is,
Assuming that the number of characters is N, the maximum value of the matching evaluation information is (N ×
4).

In step S12, it is determined for each character whether or not the entered kana character string matches the kana character string of the heading. If you disagree,
In step S13, the matching evaluation information of the non-matching character is subtracted from the maximum value of the matching evaluation information to determine the degree of matching, and the process proceeds to step S14. If they match, the process proceeds to step S14.

As an example, it is assumed that matching evaluation information such as ambiguous 4231 has been attached. And suppose that a * character string has been entered. When all the character strings match, the matching degree is 4 + 4 + 4 + 4 = 16. Here, since the last character does not match, the matching degree is:
It is calculated as 4 + 4 + 4 + 4-1 = 15.

In step S14, it is determined whether or not the match search has been completed. In other words, it is determined whether or not all characters in the kana character string to be matched are checked for a match or a mismatch.

In step S15, it is determined whether or not the input kana character string matches the kana character string of the heading. That is, if the obtained degree of matching is equal to or greater than a predetermined level with respect to the maximum value of the matching evaluation information, the input kana character string is considered to match the kana character string of the heading, and if it is less than the predetermined level. In this case, both kana character strings are regarded as mismatched. For example, assuming that a case where the degree of coincidence is 90% or more with respect to the maximum value of the coincidence evaluation information is regarded as coincidence, in the above-mentioned example “ambiguity”, 16 × 0.9 = 14.4 or more coincides. . In the case of “ambiguous” in the above-described example, since the degree of coincidence is 15, it can be considered that they coincide.

If it is determined in step S15 that the input kana character string matches the kana character string of the heading, kanji conversion is performed, the converted kanji is output, and the flow ends. . If it is determined that the input kana character string and the heading kana character string do not match, a result of “mismatch” is output and the flow ends.

According to this embodiment, each character of the heading of the dictionary unit 4 is determined by a predetermined rule and is given a value represented by four levels, that is, matching evaluation information.
The maximum value of the matching evaluation information expressed by (number of characters N × 4) is obtained, and the kana character string decomposed into the input phrase and the kana character string of the heading set in the dictionary unit 4 are represented by one character. For each case, it is determined whether or not there is a match, and only in the case of a mismatch, the match evaluation information of the unmatched character is subtracted from the maximum value of the match evaluation information. Is determined to be equal to or greater than the maximum value of, the kana character string input and decomposed into phrases matches the kana character string of the heading, and a candidate kanji corresponding to the heading is selected. .

Accordingly, an ambiguous match search can be performed. In other words, if the operator intends to input “ambiguity” but enters “aimao” and converts it to kanji,
Can be obtained.

Further, erroneous conversion of kanji due to a trivial key input mistake can be eliminated, and even if an operator performs an incorrect key input to perform kana-kanji conversion, a character string desired by the operator can be received. As a result, the trouble of correcting the input kana character string can be saved, and the work efficiency can be improved.

[0031]

According to the present invention, there is an effect that an ambiguous match search can be performed. As a result, it is possible to eliminate erroneous conversion of kanji due to trivial key input mistakes, and to receive the character string desired by the operator even if the operator performs an incorrect key input and performs kana-kanji conversion. is there. In addition, there is an effect that the labor of correcting the input kana character string can be omitted, and the working efficiency can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a kanji conversion device according to the present invention.

FIG. 2 is a diagram showing an example of headings and matching evaluation information in the kanji conversion device according to the present invention.

FIG. 3 is a flowchart showing a match search process in the kanji conversion device according to the present invention.

FIG. 4 is a block diagram showing a conventional example of a kanji conversion device.

[Explanation of symbols]

 1,31 character input unit 2,32 phrase analysis unit 3,33 search unit 4,34 dictionary unit 5,35 match detection unit

Claims (1)

(57) [Claims] 1. A character input means for inputting a character, a phrase analyzing means for decomposing an input character string into phrases to form phrase analysis information, and a character string displayed and provided with matching evaluation information. A heading, a dictionary in which at least the kanji corresponding to the heading is set as a candidate, a search means for searching a dictionary based on the sentence analysis information, and a match between the phrase analysis information and the heading of the dictionary for each character Match detecting means for detecting and subtracting the coincidence evaluation information of the non-matching character from the maximum value of the coincidence evaluation information to obtain a degree of coincidence, wherein the degree of coincidence is a predetermined level with respect to the maximum value of the coincidence evaluation information In the above, a kanji conversion device comprising a conversion means for converting the sentence analysis information into kanji.