JP3344793B2 - Kana-Kanji conversion device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a kana-kanji conversion device in a Japanese processing device.

[0002]

2. Description of the Related Art When performing kana-kanji conversion on a unit called a phrase, an independent word (noun, verb, etc.) is first searched from the beginning of the input, and a subsequent auxiliary word (particle, auxiliary verb, etc.) is searched. Above, a method of performing kana-kanji conversion by combining these search results is general. In other words, the conventional kana-kanji conversion device is based on the assumption that a character string is input in the order of an independent word and an adjunct word (there may be no adjunct word). If it is done, correct kana-kanji conversion cannot be performed.

[0003] In addition, there has been proposed a device for judging whether an attached word is valid as a head of a conversion result when an attached word, an independent word, an attached word, ... are input in this order. For example, in the “Kana-Kanji conversion method” described in Japanese Patent Application Laid-Open No. 2-2555941, after recognizing a word immediately before an input position, a connection determination is made with a conversion candidate immediately before a conversion candidate whose first word is an attached word. A method is disclosed.

In the "document creation apparatus" described in Japanese Patent Application Laid-Open No. 2-28934, a means for storing grammatical information of a phrase converted immediately before is provided. A method for performing connection determination of a certain candidate is disclosed.

FIG. 6 shows the basic configuration of the above document creation apparatus.
This will be described with reference to FIG.

The input / output control unit 61 includes a keyboard for inputting reading data and the like, and a display for sequentially displaying conversion candidates and determined kanji-mixed sentences during input reading and kanji conversion processing. The input / output control unit 61 is connected to the conversion control unit 62. The control unit 62 is configured to control processing of each unit connected thereto.

A dictionary search unit 63 searches a dictionary unit 64 that stores kanji data, kana data, point number data, and the like based on a reading instructed to be converted. The dictionary search unit 63 and the dictionary unit 64 are connected in series to the conversion control unit 62.

When the operator selects and confirms a conversion candidate, the preceding phrase determination processing unit 65 extracts the grammatical information of the immediately preceding phrase from the conversion candidate. The previous phrase buffer 66 includes a processing unit 65
The grammatical information such as the part of speech and the presence or absence of an adjunct word of the previous sentence determined in is stored. The attached word determination unit 67 extracts the grammatical information stored in the buffer 66, and determines whether at least the first part of the next input read data can be connected to the preceding phrase as an attached word. Processing unit 65 and determination unit 6
7 are connected to the control unit 62, respectively. Buffer 66
Are connected to the processing unit 65 and the determination unit 67.

[0009]

In an apparatus which is assumed to be input in the order of the independent word and the attached word, when the input operation is performed such that the independent word and the attached word are continuous, the independent word is temporarily added. In the case where the conversion is performed only with the word part, for the succeeding attached word part and below, after determining the attached word part as no conversion, further input (such as an independent word) is performed. Or, if the desired notation cannot be obtained as a result of inputting and including kana-kanji characters and performing kana-kanji conversion, a notation with the head of the attached word is obtained by correcting the conversion result using a cancel key or the like. Will be.

[0010] For example, when inputting "I went shopping with a friend.", After inputting and converting only the "Friends" portion, and inputting "Tokai." The conversion result is displayed.

This causes the user to always be conscious of inputting in the order of the independent word and the attached word, which is one of the causes of lowering the input efficiency. For this reason, it is conceivable to generate and display candidates with the adjunct word at the head, but generating these candidates unconditionally increases the number of unnecessary candidates, and also reduces the efficiency. Invite.

On the other hand, the above-mentioned prior art is disclosed. However, JP-A-2-25594
In the device described in No. 1, in order to recognize the immediately preceding word at an arbitrary input position in a sentence, it is necessary to perform a complicated process for recognition, such as a need to search a word dictionary from Kanji notation. The problem arises. Also, Japanese Patent Application Laid-Open No. 2-2
In the apparatus described in Japanese Patent No. 89034, information on a phrase that was determined immediately before (presence or absence of part-of-speech or an attached word in an independent word portion) is held for a candidate whose head is an adjunct word, and after generating a conversion candidate, Since the connection is determined based on the grammar information, the conversion processing speed is reduced.

An object of the present invention is to provide a kana-kanji conversion device capable of correctly converting a kana character string starting from an attached word into a kana-kanji sentence.

[0014]

According to the present invention, the above-mentioned objects are achieved by input means for inputting a kana character string, conversion means for performing kana-kanji conversion on the input character string, and conversion means required for the conversion. A word dictionary storage unit, a word table storage unit, a display unit for displaying a kana-kanji conversion result and the like, a succeeding possible auxiliary word table storage unit storing information on an auxiliary word that can follow the immediately preceding kana-kanji conversion result, and a kana-kanji conversion Matching means for comparing the conversion result candidate generated in the processing with the succeeding possible auxiliary word table, and updating means for updating the subsequent possible auxiliary word table based on the previously determined kana-kanji conversion result Wherein the conversion means is stored in the succeeding possible auxiliary word table storage unit,
After being updated based on the previously determined Kana-Kanji conversion result
A kana-kanji conversion is performed based on the validity of a conversion result starting from an appendix that can be succeeded on the basis of information on an appendix that can be continued. Achieved by the kana-kanji conversion device.

Further, according to the present invention, in addition to the above-mentioned object, the present invention provides, in addition to the configuration of the first apparatus, a conversion means for generating a conversion candidate for a kana-kanji conversion process by collating with a succeeding appendix table by a collation means. This is achieved by a second kana-kanji conversion device configured to generate a conversion candidate with the attached word at the head only when the attached word can follow.

[0016]

In the present invention, a succeeding appendix table for storing information of appendixes that can be used as the head of input is prepared, and based on the final phrase (or word) of the Kana-Kanji conversion result determined immediately before. Then, the succeeding appendix table is updated, and in the first apparatus of the present invention, when a conversion candidate having the appendix at the head is generated, it is used to determine the validity. Further, in the second device of the present invention, the generation of unnecessary conversion candidates starting from the adjunct word is suppressed by referring to the succeeding adjunct word table when generating the conversion candidate.

Thus, it is possible to cope with a case where the head of the input starts with an attached word, and it is possible to realize a device having a high degree of freedom with respect to the input. Further, as compared with the above-mentioned two conventional techniques, the determination of the connection validity between the head attached word and the immediately preceding conversion result can be realized by simpler processing, so that a reduction in the conversion processing speed can be suppressed as much as possible.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The basic structure of a preferred embodiment of the kana-kanji conversion device of the present invention will be described below with reference to FIG.

The input / output control unit 1 includes a keyboard for inputting reading data and the like, and a display for sequentially displaying conversion candidates and determined kanji-mixed sentences during input reading and kanji conversion processing.

The input / output control unit 1 is connected to the conversion control unit 2. The control unit 2 is configured to control processing of each unit connected thereto.

The dictionary search unit 3 searches the dictionary unit 4 storing various word dictionaries and various tables used in the kana-kanji conversion process based on the reading instructed by the conversion. The search unit 3 and the dictionary unit 4 are connected to the control unit 2 in series.

The table storage unit 5 stores a table of ancillary words that can follow an independent word. The table collation unit 6 collates a table of auxiliary words that can be stored in the storage unit 5 and can follow the independent word. The table updating unit 7 creates a table of ancillary words that can follow the independent word, based on the last phrase to be converted until immediately before, and updates the table in the storage unit 5. The collating unit 6 and the updating unit 7 are connected to the control unit 2 respectively. The storage unit 5 is connected to the matching unit 6 and the updating unit 7.

Hereinafter, a more specific configuration of the present embodiment will be described in detail with reference to FIG.

The input means 201 includes a keyboard for inputting reading data and the like, a pen input tablet and the like. The output means 202 includes a CRT display for displaying a conversion candidate and a determined kanji mixed sentence sequentially during the kanji conversion process,
D display etc. are included. The input unit 201 and the output unit 202 as a display unit are connected to an input / output control unit 203 that controls the input unit 201 and the output unit 202. The control unit 203 includes a conversion control unit 2 as a conversion unit.
04. The control unit 204 is configured to control processing of each unit connected thereto.

The dictionary search unit 205 stores a word dictionary storage unit 206 for storing a word dictionary composed of readings and notations, and a table storage unit 207 for storing various word tables used in kana-kanji conversion processing. Search based on The search unit 205 is connected to the control unit 204, and the storage units 206 and 207 are connected to the search unit 205.

The table storage unit 208 stores a table of attached words that can follow an independent word. The table collating unit 209 as a collating unit refers to a table of ancillary words that can be added to the independent word and stored in the storage unit 208. The table updating unit 210 as an updating unit creates a table of ancillary words that can follow the independent word based on the last phrase to be converted immediately before, and updates the table in the storage unit 208. The collating unit 209 and the updating unit 210 are respectively
8 is connected. The storage unit 208 is connected to the matching unit 209 and the updating unit 210.

The processing flow of the embodiment of the apparatus of the present invention will be described below with reference to FIG.

First, as the first step of the kana-kanji conversion process, an input character string is read by the input means 201 (step S1), and the dictionary search unit 205 searches the dictionary of the storage unit 206 or the table of the storage unit 207. A conversion candidate is generated and stored in the buffer 211 (Step S2). At this time, generation of a candidate having an attached word at the beginning is allowed. For example, when "I went to Tokai." Is input, candidates such as "To / to the city" and "To / to go / buy" are generated. However, the symbol / is used for convenience to indicate a segment break, and does not appear in the actual conversion result.

Next, it is checked whether or not any of the generated candidates starts with an adjunct word (step S3). If there is any, the succeeding adjunct word table (hereinafter, referred to as the following) stored in the storage unit 208 is checked. The collation unit 209 performs collation with the subsequent table. At this time, it is first checked whether or not the subsequent table can be currently collated (step S4). This is because when the subsequent table is in the initial state (at the start of input work) or when the input up to just before is not an input from a position following it (such that it is newly inserted in the middle of a previously created sentence). This is because the contents of the subsequent table cannot be used at this time. If the succeeding table is collatable in step S4, collation of the head appendix in the conversion candidate with the structure table is performed (step S4).
5).

FIG. 4 is a schematic diagram of the structure table. The table includes an ID portion for identifying each attached word, and an enable / disable bit portion indicating whether the attached word can be the head of the conversion result by 1/0 bit information. Here, for convenience, the ID portion is shown in the form of an accessory word, but in practice, a method of using a number uniquely assigned to each accessory word may be used. Alternatively, as shown in FIG. 5, the following table itself may be represented by a 1/0 bit pattern, and a mask pattern may be prepared for each attached word. As described above, when the method as shown in FIG. 4 is used, the subsequent table is searched according to the number of the head attachment word in the conversion candidate, and data indicating whether or not subsequent is possible is obtained. FIG.
In such a case, the mask data corresponding to the head adjunct word in the conversion candidate is extracted, and a logical product with the subsequent table is taken, for example, to obtain data indicating whether or not the succeeding is possible.

As a result of the collation, candidates beginning with an auxiliary word that cannot be succeeded may be deleted from the buffer 211,
Although it is not deleted as a conversion candidate, the priority may be lowered in the subsequent candidate determination processing. Step S
In the case where the collation of the succeeding table is impossible in 4, a similar treatment can be considered for the conversion candidate starting with the attached word.

The apparatus performs a final conversion candidate determination process based on the results of these processes (step S).
6). In general candidate determination processing, likelihood calculation is often performed by combining a plurality of processes using dictionary information (word frequency, grammatical connection availability, etc.) and example information. Then, details of those processes will be omitted. In addition, the comparison result of the preceding succeeding table is also taken into account in the conversion candidate determination process.

After the candidate determination processing, the conversion result that has become the first candidate is displayed on the output means 202 (step S7), and a signal indicating the determination of the conversion result (a determination signal input from the input means 201 or a new signal). It is in a state of waiting for input of a character string (step S8). During this time, in some cases, processing such as re-conversion and display of the next candidate is performed. If the signal indicates the end of the input operation (step S9), the kana-kanji conversion process ends.

When a decision signal signal other than the end signal is received, the updating unit 210 updates the subsequent table in the storage unit 208 (step S10). That is,
When the state of the final part of the conversion candidate for the immediately preceding input is determined, the subsequent table is updated (or initialized) based on the final state. Update of subsequent table
Based on the part of speech or inflected form of the last part of the last conversion candidate,
This is performed by setting a bit corresponding to an appendix that can be followed in the subsequent table as shown in FIGS. After the subsequent table is updated, the input character string is read again by the input means 201 (step S).
1) The same processing is repeated.

[0035]

According to the apparatus of the present invention, it is possible to judge the connection validity between the head attached word and the immediately preceding conversion result by a simpler process, so that a reduction in the conversion processing speed can be suppressed as much as possible.

[Brief description of the drawings]

FIG. 1 is a basic configuration diagram of a preferred embodiment of the apparatus of the present invention.

FIG. 2 is a block diagram showing a more specific configuration of one embodiment of the apparatus of the present invention.

FIG. 3 is a processing flowchart showing a processing flow of an embodiment in the apparatus of the present invention.

FIG. 4 is a schematic diagram showing a succeeding appendix table in an embodiment of the apparatus of the present invention.

FIG. 5 is a schematic diagram showing a succeeding appendix table in an embodiment of the apparatus of the present invention.

FIG. 6 is a basic configuration diagram of a conventional kana-kanji conversion device.

[Explanation of symbols]

 1, 203 input / output control unit 2, 204 conversion control unit 3, 205 dictionary search unit 5, 208 table storage unit 6, 209 table collation unit 7, 210 table update unit

Claims (2)

(57) [Claims] 1. Input means for inputting a kana character string, conversion means for performing kana-kanji conversion on the input character string,
A word dictionary storage unit and a word table storage unit required for conversion, and display means for displaying a kana-kanji conversion result and the like,
The succeeding possible supplementary word table storage unit that stores information on the succeeding supplementary words in the immediately preceding Kana-Kanji conversion result, and the conversion result candidate generated in the Kana-Kanji conversion process is compared with the succeeding possible supplementary word table. Matching means, and updating means for updating the succeeding possible supplementary word table based on the previously determined kana-kanji conversion result, wherein the conversion means is stored in the succeeding possible supplementary word table storage unit.
Is updated based on the previously determined Kana-Kanji conversion result.
Kana-Kanji characters based on the validity of the conversion result starting from the possible appendix , based on the information
Kana-kanji conversion apparatus characterized by being configured to convert. 2. A conversion method according to claim 1, wherein said conversion unit generates a conversion candidate in a kana-kanji conversion process such that the conversion unit sets the heading of the auxiliary word only when the auxiliary word can be succeeded in the collation with the succeeding auxiliary word table by the matching unit. The apparatus of claim 1, wherein the apparatus is configured to generate a candidate.