JPH1153354A - Kana/kanji converter and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a KANA (JAPANESE SYLLABARY) KANJI (Chinese character) converter performing KANA/KANJI conversion while performing processing of a homonym with a conversion object character string containing a demonstrative by using cooccurrence information. SOLUTION: A conversion means 3 generates a conversion candidate for a conversion object character string on the basis of a specified conversion rule. A cooccurrence information storage means 9 stores cooccurrence information for performing a processing of a homonym. When a given conversion object character string contains a demonstrative, an antecedent candidate decision means 11 decides the antecedent candidate indicated by a demonstrative from an established character string stored in an established character string storage means 7. A conversion means 3 replaces the demonstrative with the antecedent candidate and executes a conversion by using the cooccurrence information.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a kana-kanji conversion device and a kana-kanji conversion method, and more particularly, to a conversion process including an instruction.

[0002]

2. Description of the Related Art Japanese Unexamined Patent Publication No. Hei 5-158917 discloses a kana-kanji conversion device for performing proper conversion of homonyms using co-occurrence information.

[0003] Conversion using co-occurrence information will be briefly described. The words used in pairs are stored in the co-occurrence information dictionary in advance. For example, "tree" and "cut", "kimono"
And "wear" are stored. Then, when "cut off" is input, "cut off" is converted to "cut", and when "cut off kimono" is input, "cut off" is converted into "wear". Thereby, the homonym can be appropriately converted.

[0004] However, even when the co-occurrence information is used, if "I got a big tree" is converted to "I got a big tree" and then "It is difficult to cut it" is input. Regarding the read data "Kiru", conversion using the co-occurrence information cannot be performed.

The present invention solves the above-mentioned problem, and provides a kanji conversion apparatus and method for performing a kana-kanji conversion while processing a homonymous word using a co-occurrence information with a conversion target character string including a demonstrative. The purpose is to:

[0006]

According to a first aspect of the present invention, there is provided a kana-kanji conversion apparatus for converting a character string to be converted based on a predetermined conversion rule. If it is found, it searches for the antecedent candidate indicated by the demonstrative, replaces the definitive with the antecedent candidate, and processes the homonym using the co-occurrence information while performing the conversion. It is characterized by performing.

[0007] In the kana-kanji conversion device of claim 2,
Conversion means for generating a conversion candidate based on a predetermined conversion rule for a character string to be converted, conversion candidate storage means for storing the conversion candidate, and stored in the conversion candidate storage means upon receipt of a confirmation command from an operator. Co-occurrence information storage means for storing co-occurrence information for processing homonymous words in a kana-kanji conversion device comprising a fixed character string storage means for storing conversion candidates as fixed character strings; In the case where a descriptive term is included in the sequence, the predecessor candidate indicated by the descriptive term is selected from the conversion candidates stored in the conversion candidate storage means or the definitive character string stored in the definitive character string storage means. It is provided with means for determining antecedent candidates, and the conversion means performs the conversion using the co-occurrence information by replacing the descriptives with the antecedent candidates.

According to a third aspect of the present invention, there is provided a kana-kanji conversion device.
The conversion unit generates the conversion candidate by performing segmentation based on a predetermined segmentation rule, and when there are a plurality of segmentation methods, a plurality of conversion methods based on the plurality of segmentation methods. A conversion candidate is generated, and conversion using the co-occurrence information is performed on the plurality of conversion candidates.

According to a fourth aspect of the present invention, there is provided a kana-kanji conversion method.
In the kana-kanji conversion method of converting a character string to be converted based on a predetermined conversion rule, if a given character string to be converted includes an instructive, a search is made for the antecedent candidate indicated by the indicium. The present invention is characterized in that the conversion is performed while replacing homologous words with co-occurrence information by replacing the demonstratives with the antecedent candidates.

The storage medium according to claim 1 is a computer-readable storage medium storing a computer-executable program, and the program implements the apparatus or method according to any one of claims 1 to 4. It is characterized by that.

An embodiment realized by using the CPU 23;
The relationship with each means of each claim will be described. The conversion means 3 includes the steps ST3 and ST
This corresponds to the processing of No. 17. The conversion candidate storage means is a memory 2
7 corresponds to the conversion work buffer 27h. The determined character string storage means corresponds to the determined result buffer 27k of the memory 27. The co-occurrence information storage means corresponds to the co-occurrence example dictionary unit 26c of the hard disk 26. The antecedent candidate determining means corresponds to the processing of step ST11 of the CPU 23.

[0012]

According to the kana-kanji conversion device of the first aspect or the kana-kanji conversion method of the fourth aspect, when a given character string to be converted includes an instruction, the antecedent candidate indicated by the instruction is indicated. , And the conversion is performed while processing the homonym using the co-occurrence information, replacing the demonstrative with the antecedent candidate. Therefore, it is possible to appropriately convert the conversion target character string including the directive using the co-occurrence information.

[0013] In the kana-kanji conversion device of claim 2,
The antecedent candidate determining means, if the given character string to be converted includes an instruction, the conversion candidate stored in the conversion candidate storage means or the confirmed character stored in the confirmed character string storage means From the columns, antecedent candidates indicated by the demonstratives are determined. The conversion means replaces the demonstrative with the antecedent candidate and performs conversion using the co-occurrence information.
Therefore, it is possible to appropriately convert the conversion target character string including the directive using the co-occurrence information.

According to a third aspect of the present invention, there is provided a kana-kanji conversion device.
The conversion unit generates the conversion candidate by performing segmentation based on a predetermined segmentation rule, and when there are a plurality of segmentation methods, a plurality of conversion methods based on the plurality of segmentation methods. A conversion candidate is generated, and conversion is performed on the plurality of conversion candidates using the co-occurrence information. Therefore, even if the phrase break is not unique,
Conversion using the co-occurrence information can be performed.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS

1. Description of Functional Block Diagram An embodiment of the present invention will be described with reference to the drawings. The kana-kanji conversion device 1 shown in FIG.
It includes a conversion unit 3, a conversion candidate storage unit 5, a fixed character string storage unit 7, a co-occurrence information storage unit 9, an antecedent candidate determination unit 11, a determination unit 13, and a display unit 15.

The conversion target character string storage means 2 stores a conversion target character string which is input data input by the operator. The conversion unit 3 converts the conversion target character string based on a predetermined conversion rule to generate a conversion candidate. The conversion candidate storage unit 5 stores the conversion candidates. Co-occurrence information storage means 9
Stores co-occurrence information for processing homonyms. The antecedent candidate determination means 11 determines whether or not the given conversion target character string contains an instruction, and if the instruction is included, the finalized character string stored in the determined character string storage 7. From the character string, determine the antecedent candidate indicated by the demonstrative. Note that there may be a plurality of antecedent candidates. The conversion means 3 replaces the demonstrative with the antecedent candidate, determines whether or not conversion using the co-occurrence information is possible, and if possible, performs conversion using the co-occurrence information.

On the other hand, when the re-conversion using the co-occurrence information is not necessary, the conversion means 3 displays the conversion candidates on the display means 5. The determination unit 13 receives the determination command from the operator, determines the conversion candidates stored in the conversion candidate storage unit 5, and stores the conversion candidates in the determined character string storage unit 7 as a determined character string.

Therefore, the kana-kanji conversion device 1 searches for the antecedent from the data stored in the definite character string storage means 7 when the conversion target character string including the demonstrative is given, and uses the co-occurrence information. Conversion can be performed.

The antecedent candidate determination means 11 may search for the antecedent candidate from the conversion candidates stored in the conversion candidate storage means 5.

In the present embodiment, the antecedent candidate determining means 11 determines whether or not the given character string to be converted includes a demonstrative character. Although the antecedent candidate indicated by the demonstrator is determined from the determined character string stored in the determined character string storage means 7, the determination need not be performed. Further, the conversion means 3 replaces the denotative with the antecedent candidate,
It is determined whether or not conversion using the co-occurrence information is possible, and if possible, conversion using the co-occurrence information is performed, but without performing the determination, conversion using the co-occurrence information is performed. It may be performed.

The relationship between the embodiment realized by using the CPU 23 described below and each means shown in FIG. 1 will be described. The conversion means 3 corresponds to the processing of step ST3, step ST15, and step ST17 of the CPU 23. The conversion candidate storage means 5 corresponds to the conversion work buffer 27h of the memory 27. The confirmed character string storage means 7
7 corresponds to the determined result buffer 27k. The co-occurrence information storage means 9 stores the co-occurrence example dictionary section 26 c
Corresponds to. The antecedent candidate determining means 11 corresponds to the processing of step ST9 and step ST11 of the CPU 23.

2. 2. Hardware Configuration FIG. 2 shows an example of a hardware configuration in which the device 1 shown in FIG. 1 is realized using a CPU.

The apparatus 1 includes a CPU 23, a memory 27, a hard disk 26, a CRT 30, a FDD 25, a keyboard 28, a mouse 31, and a bus line 29.
The CPU 23 controls each unit via a bus line 29 according to a control program stored in the hard disk 26.

The control program is, for example, FDD3
3 through a flexible disk (diskette) in which the program is stored. In addition to the flexible disk, a hard disk may be installed from a computer-readable storage medium in which a program such as a CD-ROM or an IC card is substantially integrated. Furthermore, you may make it download using a communication line.

In the present embodiment, the program is installed on the hard disk 26 from the flexible disk, so that the computer indirectly executes the program stored on the flexible disk. However, the present invention is not limited to this.
Alternatively, it may be executed directly from. The programs executable by the computer include those that can be directly executed by simply installing the software, programs that need to be converted once (for example, decompressing data that has been compressed, etc.), , Which can be executed in combination with other module parts (OS, library, etc.).

The memory 27 includes a reading buffer 27y, a conversion work buffer 27h, a confirmed result buffer 27k, and a preceding word check buffer 27c. The reading buffer 27y stores the input reading character string. The conversion work buffer 27h is a work area for performing kana-kanji conversion. The determination result buffer 27k is an area for storing a determination result when the determination key is pressed.
The antecedent check buffer 27c is an area for temporarily storing candidates for the antecedent check.

The hard disk 26 stores the control program 2
6a, a basic dictionary unit 26b, a co-occurrence example dictionary unit 26c, and an auxiliary word dictionary unit 26d. Control program 26
“a” stores a control program of the apparatus, and the basic dictionary unit 26b stores a basic dictionary in kana-kanji conversion. As shown in FIG. 3, the co-occurrence example dictionary unit 26c stores co-occurrence examples for processing homonymous words in pairs. Although not shown in FIG. 3, related information such as the reading, the part of speech, and the type of use of the verb (five-step use, etc.) is also stored as in the conventional case. The auxiliary word dictionary unit 26d is a dictionary unit that stores an example of an auxiliary word (particle, etc.).

Conversion candidates and the like are displayed on the CRT 30.

3. Flowchart Next, a control program stored in the hard disk 26 will be described. In the following, a description will be given assuming that "She got a large tree." Has already been stored in the confirmed result buffer 27k.

The CPU 23 reads the input character string into the reading buffer 27y (step ST1 in FIG. 4). In the following, a description will be given assuming that the character string to be converted “difficult to cut it” is input from the keyboard 28 and read into the reading buffer 27y.

The CPU 23 separates phrases and specifies conversion candidates based on a predetermined rule for separating phrases and kana-kanji conversion rules, and stores them in the conversion work buffer 27h (step ST3). Here, as shown in FIG. 5A, the character string to be converted “it is difficult to cut it”
However, it is assumed that the conversion candidate has been converted to "difficult to wear" in consideration of the frequency of use and the like.

Next, the CPU 23 adds the co-occurrence example dictionary 26 to the conversion candidates stored in the conversion work buffer 27h.
It is determined whether or not there is any item corresponding to c (step S
T5). In this case, since the co-occurrence example of the co-occurrence example dictionary section 26c does not exist in the data “It is difficult to wear” stored in the conversion work buffer 27h, the step ST
Proceed to 9 to determine whether or not there is a directive. In the present embodiment, whether or not an instruction is present is determined based on part-of-speech information (not shown) stored for each phrase in the conversion work buffer 27h, but is not limited to this. Instead, for example, a descriptive word table that stores words to be descriptives may be prepared, and it may be determined whether or not there is a corresponding one.

If the CPU 23 determines in step ST5 that there is a co-occurrence example, it performs conversion using co-occurrence information as in the prior art, and proceeds to step ST9.

In this case, the CPU 23 determines in step ST9
In step S, since the demonstrative "it" exists,
Proceeding to T11, a candidate that can be an antecedent (hereinafter referred to as antecedent candidate) is specified from the data stored in the determination result buffer 27k and stored. In the present embodiment,
As for the antecedent candidate, a phrase including a noun phrase is selected based on the part of speech information stored in the definite buffer 27k (not shown). In this case, the CPU 23 specifies the antecedent candidates “her” and “tree” in the confirmed result buffer 27k and stores them in the antecedent check buffer 27c. This state is shown in FIG. 5B.

Next, the CPU 23 determines whether or not all extracted antecedent candidates exist in the co-occurrence example dictionary unit 26c (step ST15 in FIG. 4). In this case, first
“She” is replaced with “It”, and it is “difficult to cut her”, and it is determined whether or not it exists as a co-occurrence example in the co-occurrence example dictionary section 26c. In this case, since there is no co-occurrence example, it is determined whether or not a co-occurrence example exists in which “the tree is hard to cut” is replaced by “the tree”, which is the next candidate. In this case, since "tree" and "cut" exist as co-occurrence examples in the co-occurrence example dictionary section 26c, re-conversion processing is performed (step ST17). Thus, as shown in FIG. 5C, the conversion work buffer 27h stores the conversion candidate “It is difficult to cut it”. The CPU 23 displays the conversion candidate “It is difficult to cut it” stored on the conversion work buffer 27h on the CRT 30 (FIG. 4, step ST19).

The operator looks at the displayed conversion candidates and operates the enter key or the cancel key on the keyboard 28. C
The PU 23 determines whether or not the enter key or the cancel key is operated (step ST21). If the enter key is pressed, it is determined (step ST2).
3). Specifically, the CPU 23
The data of the conversion work buffer 27h is added to k and stored. On the other hand, if the cancel key is pressed in step ST21, the process ends.

As described above, in the present embodiment, for the antecedent candidate that can be the antecedent, the determined result buffer 2
7k is searched, all antecedent candidates are extracted, and it is determined whether or not there is a match with the co-occurrence example stored in the co-occurrence example dictionary unit 26c. When there is an antecedent candidate, the conversion candidate is replaced with the word of the co-occurrence example and converted. This allows
Even when there is a directive, conversion with higher conversion efficiency can be performed.

Although a plurality of antecedent candidates are extracted when there are a plurality of antecedent candidates, priority rules may be stored to narrow the antecedent candidates. For example, among the noun phrases, the declarative is "it",
If it is "this", it is difficult to refer to "people"
"She" may be excluded from the antecedent candidates. In other words, the properties of the demonstratives (whether they are human or physical descriptives, etc.)
And the like may be stored, and narrowed down according to the property.

It is to be noted that instead of searching all the converted character strings stored in the determined result buffer, a limited range such as only one sentence or two sentences separated by a period (.) Is searched. Is also good.

In this embodiment, antecedent candidates are extracted from the determined result buffer 27k. With this, even if you input a fragment for each phrase,
Phrases including demonstrators can be appropriately converted. However, the present invention is not limited to this, and the conversion work buffer 27h may be searched. Thus, even when two readings are input and conversion is performed at a time, a phrase including a directive can be appropriately converted.

In the above embodiment, the homonym conversion using the co-occurrence example is performed in the same manner as in the prior art.
Although the search was performed using the reading and the part of speech as keys, the present invention is not limited to this method, and other methods can be applied.

When the input character string is divided into phrases,
When there are a plurality of delimiters, each of them may be stored in the conversion work buffer 27h, and the antecedent candidate may be checked for each.

In the present embodiment, conversion is performed once using an ordinary basic dictionary, and if there is a co-occurrence example, it is re-converted using this and displayed as a conversion candidate, and there is no co-occurrence example. In this case, it is displayed as a conversion candidate as it is, and waits for the operator to input a confirmation command. Therefore,
The conversion candidate is displayed without the operator being conscious of whether or not the re-conversion has been performed.

In this embodiment, the case where the present invention is applied to a Japanese word processor has been described. However, the present invention can be used as a FEP of a Japanese word processor. Furthermore, the present invention is also applied to a case where a kana-kanji conversion device is used as an application program other than the Japanese word processor, for example, a FEP such as a CAD program or a spreadsheet program.

In the present embodiment, the functions shown in FIG. 1 are realized by using the CPU 23 and software. However, some or all of them may be realized by hardware such as a logic circuit.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of Kana-Kanji conversion 1 according to the present invention.

FIG. 2 is a diagram showing an example of a hardware configuration of the kana-kanji conversion device 1 shown in FIG.

FIG. 3 is a diagram showing a co-occurrence example dictionary.

FIG. 4 is a flowchart illustrating conversion when there is an antecedent.

FIG. 5 is a diagram showing the transition of each buffer.

[Description of Signs] 3... Conversion means 5... Conversion candidate storage means 7... 13: Individual information adding means 23: CPU 27: Memory

Claims (5)

[Claims] 1. A kana-kanji conversion device for converting a character string to be converted based on a predetermined conversion rule, wherein when a given character string to be converted includes an instruction, the antecedent indicated by the instruction is used. Kana-Kanji conversion apparatus characterized by: searching for candidates, replacing said demonstratives with said antecedent candidates, and performing said conversion while processing homonyms using co-occurrence information. 2. A conversion means for generating a conversion candidate based on a predetermined conversion rule for a character string to be converted, and a conversion candidate storage means for storing the conversion candidate. Co-occurrence information storage for storing co-occurrence information for processing homonymous words in a kana-kanji conversion device comprising: a fixed character string storage means for storing the conversion candidates stored in the conversion candidate storage means as a fixed character string. Means, when the given conversion target character string includes an instruction, from the conversion candidates stored in the conversion candidate storage means or the fixed character string stored in the fixed character string storage means, An antecedent candidate determining means for determining an antecedent candidate indicated by a demonstrative, wherein the converting means replaces the definitive with the antecedent candidate and performs conversion using the co-occurrence information. Kana-kanji conversion device that. 3. The kana-kanji conversion device according to claim 2, wherein the conversion means generates the conversion candidates by performing segmentation based on a predetermined segmentation rule, and there are a plurality of ways of segmenting the phrase. In this case, a plurality of conversion candidates are generated based on the plurality of delimiters, and the plurality of conversion candidates are converted using the co-occurrence information. 4. In a kana-kanji conversion method for converting a character string to be converted based on a predetermined conversion rule, when a given character string to be converted includes a directive, the antecedent indicated by the directive is indicated. Retrieving candidates, replacing said descriptives with said antecedent candidates, and performing said conversion while processing homonyms using co-occurrence information. 5. A computer-readable storage medium storing a computer-executable program, wherein the program implements the apparatus or method according to claim 1. Characteristic storage medium.