JP3466669B2 - Character processing method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character processing method.

[0002]

[Prior art]

(1) As shown in FIG. 2, in the conventional phrase candidate registration process of the kana-kanji conversion process, first, in step 21, the independent part information is extracted from the processing independent part information group. Then, in step 23, the retrieved independent part information is registered in the independent part information storage area of the clause table.

In step 22, when all the independent part information is extracted from the processing target independent part information group, the process proceeds to the next step. At 24, information such as the independent part information registered in the independent part information storage area of the phrase table is set as the phrase candidate creation information in order to create the phrase candidate. In 26,
The phrase candidates are created based on the phrase candidate creation information set in 24.

In step 27, the result of creating a bunsetsu candidate is judged,
If the bunsetsu candidate can be created, the process proceeds to the next step. If it cannot be created, the process proceeds to 24 because the next independent part information is processed. At 28, the created phrase candidate is registered in the phrase candidate storage area of the phrase table. In 25, when the phrase candidate creation process is performed for all the independent part information registered in the independent part information storage area of the phrase table, this process ends.

(2) As shown in FIG. 7, in a conventional phrase candidate registration process of another kana-kanji conversion process, first, a part-of-speech table is created from the extracted independent part information, and a phrase candidate is created based on the table. It was created and registered in the phrase candidate registration area of the phrase table. At the same time, the independent club information was also registered in the independent club information registration area of the clause table.

(3) Conventionally, in the phrase candidate registration process of the kana-kanji conversion process, first, if there is a phrase candidate table as shown in FIG. 12, the phrase candidates stored in the phrase candidate buffer are registered, the shortest result is obtained. The phrase candidate is deleted, and FIG.
Is registered as.

[0007]

In the above-mentioned conventional character processing method (1), all the independent part information of the independent part information group to be processed is registered in the independent part information storage area of the clause table. Therefore, since the independent section information for which the phrase candidate could not be created is also registered in the independent section information storage area of the clause table, useless independent section information is also registered. There is a problem in that the speed of the candidate registration process is reduced.

In the above-described conventional character processing method (2), the phrase candidate creating process is performed for all created part-of-speech tables, but the same part-of-speech table may be created. However, since the bunsetsu candidate creation process is performed, the same process is repeated, which causes a problem that the bunsetsu candidate registration process of the kana-kanji conversion process is slowed down.

Further, the above-mentioned conventional character processing method (3)
Then, if a punctuation mark or a conversion start code is connected to a phrase, that phrase is converted with priority, and other phrase candidates are excluded from the conversion. There was a problem that they were registered in the phrase candidate table together.

Therefore, an object of the present invention is to provide a character processing method that solves the above problems.

[0011]

In order to achieve the above object, the present invention is a character processing method executed by an information processing system including a memory in which a processing target independent section information group and a clause table are stored. The step of extracting independent part information from the processing target independent part information group on the memory, the step of creating a bunsetsu candidate based on the extracted independent part information, the created bunsetsu candidate and the bunsetsu candidate concerned can be created. And registering the independent part information in the phrase candidate storage area and the independent part information storage area of the phrase table on the memory, respectively.

Further, the present invention is a character processing method which is executed by an information processing system having a phrase candidate table and a memory storing a conversion priority phrase length data, and having a phrase candidate buffer. Based on the phrase length data, the phrase candidates in the phrase candidate buffer can be registered in the target phrase registerable determination step that determines whether the phrase candidates can be registered in the phrase candidate table on the memory, and in the target phrase registerable determination step. A registration step of registering the determined phrase candidate in the phrase candidate table on the memory.

[0013]

[0014]

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

<Embodiment 1> Next, Embodiment 1 of the present invention will be described in detail with reference to the drawings.

FIG. 3 is a block diagram showing the arrangement of an information processing system for executing the character processing method common to the respective embodiments of the present invention. In FIG. 3, reference numeral 1 denotes a central processing unit that executes character processing including kana-kanji conversion processing, and RO
The character processing procedure as shown in FIGS. 1, 6 and 11 stored in M2 is executed. The ROM 2 is shown in FIGS.
It stores each information used when executing the character processing procedure shown in. The display device 3 outputs the processing result of the kana-kanji conversion processing and the like. The input device 4 such as a keyboard is used to input a reading string or the like. The RAM 5 has a work area of the central processing unit 1, that is, an area for temporarily storing information such as clause data.

First, the basic processing of the first embodiment will be described with reference to FIG.

In the phrase candidate registration process of the kana-kanji conversion process in this embodiment, first, the processing target independent part information group created by the independent part information extraction process of the kana-kanji conversion process is stored in the RAM 5. . Take out the information of the independent section from there, and based on the information such as the information of the independent section,
Performs phrase segment creation processing. Once you have created a phrase candidate,
The bunsetsu candidate and the independence part information of which the bunsetsu candidate of the extracted independence part information can be created are registered in the bunsetsu candidate storage area and the independence part information storage area of the bunsetsu table.

FIG. 1 is a flow chart showing the processing of this embodiment.

First, in step 11, the independent section information is extracted from the processing target independent section information group. In 12, when all the independent part information is extracted from the processing target independent part information group, the process proceeds to the next step.

In step 13, information such as the extracted independent section information is set in the phrase candidate creating information in order to create the phrase candidate.

At 14, the phrase candidates are created based on the phrase candidate creation information set at 13.

In step 15, the bunsetsu candidate creation result is judged,
If the phrase candidate can be created, the process proceeds to the next step. If it cannot be created, the process returns to 11 to process the next independent part information.

Then, in 16 the self-supporting part information for which the clause candidate has been created is registered in the self-supporting part information storage area of the clause table, and in 17 the created clause candidate is registered in the clause candidate storage area of the clause table. .

FIG. 5 is a diagram showing the structure of the phrase table.

The phrase table is composed of a phrase candidate storage area, an independent section information storage area, and other information.

As described above, according to the first embodiment, only the independent part information for which the bunsetsu candidate can be created can be registered in the independent part information storage area of the bunsetsu data. Since information registration can be omitted, there is an effect that the phrase candidate registration process of the kana-kanji conversion process can be speeded up. Further, since the number of registrations is reduced, there is an effect that the storage area is reduced.

<Second Embodiment> First, the basic processing of the second embodiment will be described with reference to FIG.

In the phrase candidate registration process of the kana-kanji conversion process in the present embodiment, first, a part-of-speech table is created based on the independent-part information extracted and extracted from the processing target by the independent-part information extraction process or the like. Check whether the same part is registered in the effective part-of-speech table and the invalid part-of-speech table in the table storage area. If the same part is registered, it means that the part-of-speech tables are duplicated.
The process ends.

Based on the part-of-speech table, a bunsetsu candidate is created. If the bunsetsu candidate can be created, it is stored in the effective part-of-speech table storage area, and if it cannot be created, it is stored in the invalid part-of-speech table storage area. To register. Then, if the phrase candidate cannot be created, the process is terminated, and if it is created, the process shifts to the next process.

The independent part information of the created phrase candidate is registered in the independent part information storage area of the phrase table, and the phrase candidate is registered in the phrase candidate storage area of the phrase table.

FIG. 6 is a flow chart showing the processing of this embodiment.

In step 31, the independent part information extracted by the independent part information extracting process or the like is extracted from the processing target. In 32,
In step 31, a determination is made when the independent part information could not be extracted, and in that case, this processing ends.

At 33, a part-of-speech table is created on the basis of the extracted independent part information. (The part-of-speech table stores detailed information representing the independent part information.) At 34, the part-of-speech table created at 33 and the effective part-of-speech table in the part-of-speech table storage area and the ineffective part-of-speech table are stored. The stored part-of-speech table is compared to determine whether or not there is the same one.

At 35, a determination is made when the same part-of-speech table is registered, and in this case, this processing ends.

At 36, a phrase candidate is created based on the created part-of-speech table.

At 37, the result of the bunsetsu candidate is determined, and if it is created, the process proceeds to step 38, and if it is not created, the process proceeds to step 41.

At 38, the part-of-speech table of the created clause candidates is registered in the effective part-of-speech table storage area of the part-of-speech table storage area.

Then, in 39, the independent section information of the created clause candidates is registered in the independent section information storage area of the clause table, and in 40, the prepared clause candidates are registered in the clause candidate storage area of the clause table.

At 41, the part-of-speech table of the phrase candidates that could not be created is registered in the invalid part-of-speech table storage area of the part-of-speech table storage area.

FIG. 9 is a diagram showing the structure of the part-of-speech table storage area. The part-of-speech table storage area is composed of information such as an effective part-of-speech table storage area and an invalid part-of-speech table storage area.

FIG. 10 is a diagram showing the structure of the phrase table. The clause table is made up of information such as a clause candidate storage area and an independent section information storage area.

As described above, according to the second embodiment, duplicate registration of the part-of-speech table can be avoided, so that the number of bunsetsu candidate creation processes can be reduced, and the bunsetsu candidate registration process of the kana-kanji conversion process can be speeded up. The effect is that you can do it.

<Third Embodiment> First, the basic processing in the phrase candidate registration processing of the kana-kanji conversion processing of the third embodiment will be described with reference to FIGS.

FIGS. 14 to 17 show the case of registering the conversion priority clause first.

In FIG. 14, there is a bunsetsu candidate created by the bunsetsu candidate creating process of the kana-kanji conversion process, and the bunsetsu candidate is stored in the bunsetsu candidate buffer. In addition, there is a conversion priority clause length that stores an invalid value at the start of processing. Then, there is a phrase candidate table in the process of being created.

A case where the conversion priority clause is stored in the clause candidate buffer in such a state will be described as an example. (The conversion priority clause is a clause in which a punctuation mark, a conversion start code, and the like follow at the end of the sentence.) In FIG. 15, in the state where an invalid value is stored in the conversion priority clause length,
When the phrase candidate in the phrase candidate buffer is a conversion priority phrase, the phrase length of the phrase candidate is calculated and registered as the conversion priority phrase length.

In FIG. 16, when the conversion priority clause length is registered, the clause candidate table is searched, all clauses shorter than the conversion priority clause length are deleted, and invalid clause candidates are omitted from the clause candidate table.

In FIG. 17, the phrase candidates stored in the phrase candidate buffer are registered in the phrase candidate table in the phrase candidate registration process.

FIGS. 18 to 19 show the case where the conversion-preferred clause is already registered in the clause candidate table and the conversion-preferred clause is not registered.

In FIG. 18, there is a bunsetsu candidate created by a bunsetsu candidate creating process of the kana-kanji conversion process, and the bunsetsu candidate is stored in the bunsetsu candidate buffer. The conversion priority phrase length stores the phrase length of the conversion priority phrase registered in the phrase candidate table. Then, there is a clause candidate table in which conversion priority clauses are registered.

The case where the conversion-priority external clause is stored in the clause candidate buffer in such a state will be described as an example. (The conversion-priority external clause is a normal clause in which punctuation marks, conversion start codes, etc. do not follow the sentence end.) In FIG. 19, the conversion-priority clause length stores the clause length of the conversion-priority clause. Since the phrase length of the phrase candidate in the phrase candidate buffer is shorter than the conversion priority phrase, it is determined that the phrase cannot be registered, and the process ends without performing the phrase registration process.

20 to 22 show the case where the conversion-preferred clause is registered while the conversion-preferred clause is not registered.

In FIG. 20, there is a bunsetsu candidate created by the bunsetsu candidate creating process of the kana-kanji conversion process, and the bunsetsu candidate is stored in the bunsetsu candidate buffer. Further, since the conversion priority clause is not registered in the clause candidate table, an invalid value is stored in the conversion priority clause length. Then, there is a phrase candidate table in which the phrase candidates are registered.

The case where the conversion-preferred external clause is stored in the clause candidate buffer in such a state will be described as an example.

In FIG. 21, since an invalid value is registered in the conversion priority clause length, it is determined that the conversion priority external clause in the clause candidate buffer can be registered in the clause candidate table.

In FIG. 22, the phrase candidates stored in the phrase candidate buffer are registered in the phrase candidate table in the phrase candidate registration processing.

FIG. 11 is a flow chart showing the processing of this embodiment.

First, at 51, it is determined whether registration is possible in the phrase candidate table based on the condition of the phrase candidates in the phrase candidate buffer and the registration condition of the conversion priority phrase length.

At 52, if the result of the determination is that the phrase candidate in the phrase candidate buffer can be registered, the process proceeds to the next step, and if it cannot be registered, this processing ends.

At 53, if the phrase candidate in the phrase candidate buffer is a conversion priority phrase, the process proceeds to the next step, and if not, the process proceeds to 60.

At 54, it is judged whether or not the conversion priority phrase is registered in the phrase candidate table. Whether the conversion priority phrase length is set to the phrase length of the conversion priority phrase or an invalid value is set. Done in.

At 55, if it is determined that the conversion priority phrase is not registered in the phrase candidate table, the process proceeds to the next step, and if not, the process proceeds to 60.

At 56, the phrase length of the conversion priority phrase is obtained,
Register to the conversion priority phrase length.

At 57, it is judged whether or not the phrase candidates having the conversion priority phrase length or less are registered in the phrase candidate table.

At 58, if a phrase candidate having a conversion priority phrase length or less is registered in the phrase candidate table, the process proceeds to the next step, and if not, the process proceeds to 60.

At 59, all the phrase candidates whose length is equal to or shorter than the conversion priority phrase length registered in the phrase candidate table are deleted.

Then, at 60, the phrase registration processing is performed for the phrase candidates registered in the phrase candidate buffer in the phrase candidate table.

As described above, according to the third embodiment,
If a conversion-preferred phrase exists, only that phrase is registered in the phrase candidate table, and the process of registering non-conversion-preferred phrases can be omitted, and the phrase candidate registration process for Kana-Kanji conversion processing can be speeded up. The effect is that it can be achieved. Further, since no unnecessary phrase candidates are registered for the process of referring to the created phrase candidate table,
There is also an effect that the process of referring to the phrase candidate table can be speeded up.

[Brief description of drawings]

FIG. 1 is a flowchart showing a process of a first embodiment.

FIG. 2 is a flowchart showing the processing of a conventional technique.

FIG. 3 is a block diagram showing a configuration of an information processing system in which character processing of the present invention is performed.

FIG. 4 is a diagram illustrating a basic process of the first embodiment.

FIG. 5 is a diagram showing a phrase table storage area.

FIG. 6 is a flowchart showing the processing of the second embodiment.

FIG. 7 is a diagram showing a process of a conventional technique.

FIG. 8 is a diagram illustrating a basic process of the second embodiment.

FIG. 9 is a diagram showing a part-of-speech table storage area.

FIG. 10 is a diagram showing a phrase table storage area.

FIG. 11 is a flowchart showing the processing of the third embodiment.

FIG. 12 is a flowchart showing the processing of a conventional technique.

FIG. 13 is another flow chart representing the processing of the prior art.

FIG. 14 is a diagram illustrating a part of the basic processing of the third embodiment.

FIG. 15 is a diagram illustrating another part of the basic processing of the third embodiment.

FIG. 16 is a diagram illustrating still another part of the basic processing according to the third embodiment.

FIG. 17 is a diagram illustrating still another part of the basic processing according to the third embodiment.

FIG. 18 is a diagram illustrating still another part of the basic processing according to the third embodiment.

FIG. 19 is a diagram illustrating still another part of the basic processing according to the third embodiment.

FIG. 20 is a diagram illustrating still another part of the basic processing according to the third embodiment.

FIG. 21 is a diagram illustrating still another part of the basic processing according to the third embodiment.

FIG. 22 is a diagram illustrating still another part of the basic processing according to the third embodiment.

[Explanation of symbols]

1 Central processing unit 2 ROM 3 display devices 4 input device 5 RAM

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06F 17/22 514 JISST file (JOIS)

Claims (3)

(57) [Claims] 1. A processing target independent section information group and a clause table.
By an information processing system equipped with a memory that stores
A method for performing character processing, comprising: extracting independence part information from a processing target independence part information group on the memory; creating a phrase candidate based on the extracted independence part information; The character processing method further comprises: registering the phrase candidate and the independent part information in which the phrase candidate has been created in the phrase candidate storage area and the independent part information storage area of the phrase table on the memory , respectively. 2. A phrase candidate table and a conversion priority phrase length data.
It has a memory that stores data and has a clause candidate buffer.
Character processing method executed by the information processing system
There, based on the conversion priority clause length data on the memory, phrase candidate of the phrase candidate in the buffer, and the target phrase registrable judgment step of judging whether registerable in clause candidate table on the memory, the target Can be registered in the phrase registration possibility judgment step
The determined phrase candidates are stored in the phrase candidate table on the memory.
A character processing method comprising: a registration step for registering with . 3. The target clause registration according to claim 2 ,
If it is possible to register at the Noh judgment step, the registration is possible
Judge whether the phrase candidate determined to be a conversion priority phrase
A conversion priority clause determining step, a conversion priority clause registered determination step for determining whether a conversion priority clause is registered in the clause candidate table, and a conversion priority clause determining step for determining the conversion priority clause, Transformation priority sentence
When unregistered determination in the section registered determination step, conversion priority to be set to the memory length of the determined phrase candidates to be the conversion priority clause, as the conversion priority phrase length data
A phrase length setting step, if the length following clause is registered in the phrase candidate table, cutting deletes the phrase candidates registered <br/> Rokusumi follows the length from the phrase candidate table
And a step of removing the character.