JPH08153101A - Proofreading method for japanese sentence - Google Patents

Abstract

PURPOSE: To support proofreading by automatic checking as to errors of various chapter titles. CONSTITUTION: A morpheme analysis of a document is taken (S1) and when a specified proofread item of this document is a chapter line (S2), the chapter title part noted in the document is recognized and extracted from the morpheme analytic data (S3), and the pattern of this chapter title is stored in a tree structure (S4), whether or not there is an error in chapter title notation is checked based on the chapter title data in the tree structure (S5), and when there is a error, the chapter title is indicated to an operator to correct the chapter title by correction input by the operator (S6 and S7). To recognize the chapter title (S3), basic components noting the chapter title are prepared and the chapter title appearing in the document is recognized as a chapter title pattern which is a combination of basic components.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a Japanese language processing system, and more particularly to a method for proofreading chapter titles of documents.

[0002]

2. Description of the Related Art Japanese language processing using computers such as word processors, machine translations, document databases, and hypertexts has been put into practical use.

In natural language analysis for this purpose, a sentence to be analyzed is first divided into morpheme units (minimum units of word structure), and morpheme analysis is performed to clarify the properties of each morpheme. After this, a syntactic analysis that analyzes from the syntactic rules of natural language, and then a semantic analysis that removes ambiguity and vagueness,
Perform contextual analysis.

In such Japanese processing, as a document proofreading support system, a function for extracting / correcting typographical errors, omissions, erroneous terms and unregistered words is provided.

However, no chapter title has a proofreading function. This is because there are so many variations in how to write chapter titles that it is impossible to support all of them.

[0006]

[Problems to be Solved by the Invention] In the conventional Japanese language processing system, since the grammar proofing support system does not have a chapter title proofreading support function, manual proofreading is required, and it takes time to overlook a mistake in the chapter title. May cause

An object of the present invention is to provide a proofreading method for assisting the proofreading by automatically checking various chapter title errors.

[0008]

In order to solve the above problems, the present invention recognizes and extracts each chapter title portion written in a document, and saves the extracted chapter title patterns in a tree structure. It is characterized in that the chapter title data of the tree structure is checked for an error in the chapter title notation, the operator is instructed to specify the incorrect chapter title, and the chapter title is corrected by the operator's correction input.

Further, the present invention is characterized in that the chapter title is recognized by preparing a basic part for indicating the chapter title and recognizing the chapter title appearing in the document as a chapter title pattern which is a combination of the basic parts. To do.

[0010]

[Function] The notation error can be automatically checked by extracting the chapter title part and by making the title pattern into a tree structure.

Further, in recognition of a chapter title, recognition is performed as a combination of basic parts, so that reliable recognition can be obtained using a small amount of data.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 1 is a processing procedure diagram showing an embodiment of the present invention.

(S1) Morphological analysis is performed on an original sentence to be processed in Japanese.

(S2) The morpheme data which is the morpheme analysis result is read in, and the operator designates a calibration item (chapter title, typographical error, omission, erroneous term, etc.)

(S3) If the designated proofreading item is a chapter title, the morpheme data is used to identify the chapter title in the document.
Recognize from the beginning of the document. This recognition is performed, for example, in FIG.
As shown in the document structure and chapter title only, "..."
Only the chapter title is extracted from the document including the sentence shown in.

(S4) The recognized chapter title pattern data is stored in a tree structure. This tree structure is shown in Figure 2.
As shown in FIG. 3, the data structure for the example document has a data structure that branches into chapters, sections, subsections, and the like in this order.

(S5) It is checked from the chapter title data of the tree structure whether there is an error in the chapter title writing method. This check checks for missing permutations of chapter title numbers, duplication, and differences in notation.

For example, in the tree structure of FIG.
There is a jump from chapter to chapter 4 (missing chapter 3), and there is a difference in the notation of ◯ and ●, so check these.

(S6) The result of whether there is an error in the chapter title check and the error portion are extracted.

(S7) If there is an error, the error portion is displayed on the display device and corrected by the operator's correction input.

Therefore, according to the present embodiment, a mistake in writing the chapter title can be automatically recognized and extracted, and correction can be easily performed by the operator's correction by instructing the mistaken portion.

Next, a method of recognizing a chapter title in the process (S3) will be described. Although there are many variations in how to write chapter titles, only these basic parts are prepared for recognition, and the chapter title pattern is recognized according to the chapter title of the actual sentence.

For example, in the notation method of chapter titles, the basic parts shown in the following table are used, and these basic parts are prepared.

[0024]

[Table 1]

When actually recognizing, the following title, "3.11 Statistical data and its utilization", refers to the basic parts and says "First there are two tabs, then Is followed by a square, a number string is followed by a period, a blank is left in the number string, and a character string follows at the end. "

Characters continue from the beginning or blank 1
Those that are followed by a character string are recognized as normal sentences, not as chapter titles, and are excluded from chapter title recognition targets.

The chapter title recognition result as described above is similarly executed for the chapter titles appearing later, and the processing S
It is also used as the title pattern of No. 4, and is also used for the chapter title check of process S5.

Therefore, according to the chapter title recognition method of the present embodiment, the chapter title can be recognized only by preparing the basic parts corresponding to various patterns of writing. It can be used to create and check the tree structure of the title pattern because it becomes the identification of the pattern.

In the embodiment, when the means for separating the chapter title part from other sentences directly from the document (for example, means for separating the mixed numbers and the number of blank spaces) is used, the morphological analysis data is not always required. do not need.

[0030]

As described above, according to the present invention, the chapter title portions written in the document are respectively recognized and extracted, the pattern of the chapter titles is saved as a tree structure, and the chapter title data of the tree structure is extracted. By checking whether there is an error in the chapter title notation and instructing the operator with the incorrect chapter title and correcting the chapter title by the operator's correction input, the notation error of the chapter title is automatically checked and corrected. Therefore, the burden on the operator is reduced, and errors are not overlooked.

Further, according to the present invention, since the chapter titles are recognized by preparing the basic parts for indicating the chapter titles and recognizing the chapter titles appearing in the document as the chapter title pattern which is a combination of the basic parts, many There is an effect that automatic recognition of chapter titles that are variations of can be surely executed using a small amount of data.

[Brief description of drawings]

FIG. 1 is a processing procedure diagram showing an embodiment of the present invention.

FIG. 2 is an example of a chapter title in the document structure.

FIG. 3 is an example of a tree structure of chapter titles.

Claims (2)

[Claims] 1. A chapter title portion described in a document is recognized and extracted, and the extracted chapter title pattern is saved as a tree structure, and whether there is an error in the chapter title notation from the chapter title data of the tree structure. Is checked, the chapter title having an error is indicated to the operator, and the chapter title is corrected by the operator's correction input. 2. The chapter title recognition is characterized in that a basic part for indicating a chapter title is prepared, and the chapter title appearing in a document is recognized as a chapter title pattern which is a combination of the basic parts. How to proofread the written Japanese sentence.