JPH03252859A - Index automatic generation device - Google Patents

Abstract

PURPOSE:To generate the index of the sentence of an electronic text file in which no specified symbol is inserted automatically and at high speed by analyzing the feature of a character string to be an index item candidate. CONSTITUTION:Each character string of an index item candidate table is morpheme-analyzed by a morpheme analyzing part 3, and the character string whose word at the end of the sentence is a noun or a suffix among the morpheme-analyzed character strings is extracted by a sentence end word analyzing part 4, and is stored in an index item table. The character string of this index item table is classified by an index classifying part 5 into a corresponding large, medium and small items according to what of the large item, the medium item and the small item determined beforehand the character of the head of the sentence of each character string coincides with, and the index table is generated. Thus, the index of the sentence of the electronic text file in which the specified symbol to be recognized as the index item is not inserted in the sentence can be generated automatically.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an automatic table of contents generation device that receives a text converted into an electronic text file and automatically generates a table of contents for the text.

"Conventional technology" An example of the operation of a conventional device of this type is shown in FIG.

In the conventional technology, a specific symbol "*" is manually added to the text of the electronic text file 11 at the end of a character string that is recognized as a table of contents item, and this specific symbol is added to the table of contents item extraction unit 12 of the computer. The table of contents items are extracted by searching, the extracted table of contents items are stored in the table of contents table 13, and further displayed as a table of contents on the display 14.

``Problem to be solved by the invention'' With this conventional table of contents generation device, it is necessary to be aware of the items that will become the table of contents when creating a document, and the table of contents items in the document must be manually pointed out in advance, which is an extra task. was necessary.

An object of the present invention is to provide an automatic table of contents generation device that can automatically generate a table of contents of a text of an electronic text file in which no specific symbol recognized as a table of contents item is inserted into the text.

"Means for Solving the Problem" According to the present invention, an digitized text file is received by the original text input section, and the accepted digitized text file is processed from the beginning of the sentence to the next line break by the table of contents item candidate word extraction section. , in each character string from a line break to the next line break, or from a line break to the end of the sentence, the number of characters must be a specified value, for example 35 or less A string of characters enclosed in parentheses, etc. that is represented by Arabic numerals, Roman numerals, Chinese numerals, or symbols is extracted, and the extracted string is morphologically analyzed by the morphological analysis section.As a result of the analysis, the character string is A character string in which the word at the end of a sentence is a noun or a suffix is extracted by the sentence-final word analysis unit, and the table of contents classification unit identifies whether the first character of the sentence is an Arabic numeral, Roman numeral, or Chinese numeral. , or Arabic numerals, Roman numerals, or Chinese numerals enclosed in square brackets or square brackets,
The characters are classified into large items, medium items, and small items based on whether they are symbols or symbols, and the classified character strings are displayed as a table of contents in the table of contents generation result display section.

"Embodiment" FIG. 1 shows an embodiment of the present invention. Reference numeral 1 denotes an original text input unit that inputs a digitized text file, 2 a table of contents item candidate extraction unit that extracts character strings that are table of contents item candidates from the input digitized text file, and 3 a table of contents item candidate extraction unit that extracts character strings that are table of contents item candidates from the input digitized text file. 4 is a morphological analysis unit that performs morphological analysis on a plurality of character strings, and 4 is a sentence-final unit that analyzes the part of speech of the word at the end of a sentence in the morphologically analyzed character string and extracts a character string that is determined to be a table of contents item. A word analysis unit 5 is a table of contents classification unit that classifies the character strings extracted by the sentence final word analysis unit 4 into large items, medium items, and small items and generates a table of contents; 6 is a table of contents classification unit that displays the generated table of contents on a display. This is the table of contents generation result display section that is output to.

The operation of this embodiment will be explained with reference to FIG.

First, the electronic text file 7 is input to the original text input section l. Unlike file 11 in FIG. 4, this electronic text file 7 does not have a specific symbol "*" at the end of the character string that is recognized as a table of contents item.This electronic text file 7 is a table of contents item candidate. In extraction part 2, the number of characters is 35 for each character string from the beginning of a sentence to the next line feed, from a line feed to the next line feed, or from a line feed to the end of a sentence.
Character strings that are within the range of characters, have no period in the sentence, and in which the first character of the sentence starts with a number or symbol are extracted and stored in the table of contents item candidate table 8 as shown in FIG.

When the first character of this sentence is a number, it means an Arabic numeral,
Roman numerals, Chinese numerals, or these numbers in parentheses such as circles or keys. Further, as symbols, FIG. 2 shows examples of "O" and ".".

Next, each character string in this table of contents item candidate table 8 is morphologically analyzed by the morphological analysis section 3. In the morphologically analyzed character string, character strings in which the final word of the sentence is a noun or a suffix are extracted by the sentence final word analysis unit 4 and are extracted from the table of contents item table 9.
is stored in The character strings in the table of contents item table 9 are determined by the table of contents classification unit 5, and the first character of each character string is determined to be one of a predetermined large item, medium item, and small item as shown in FIG. 3, for example. The table of contents table 10 is generated by classifying the items into large, medium, and small items depending on whether they match.

In the example in Figure 3, the numbers in parentheses are large items, the numbers in parentheses are medium items, and the symbols are small items. Therefore, in the example in Figure 2, the large items... "1. Greetings", “2. Explanation of off-talk communication service” Middle items: “(1) Service overview and benefits”, “(
2) Regarding fees” Sub-items: “○ Service overview” and “O Benefits”.

The table of contents table 10 generated in this way is displayed on a screen by the table of contents generation result display section 6 on the display.

"Effects of the Invention" As explained above, the automatic table of contents generation device of the present invention is capable of adding specific symbols etc. to character strings that are manually recognized as being around the table of contents in advance, unlike the prior art. Unlike the method in which a computer searches and extracts it, by analyzing the characteristics of the character string that becomes a candidate for the table of contents, we find that the character string is short, the first character of the sentence is a number or symbol, and the last character of the sentence is a noun. Since the table of contents is extracted based on whether it is a suffix or a suffix, a table of contents can be generated quickly and automatically.

This eliminates the need to manually extract character strings that will become table of contents items in advance, which was done with conventional technology, and also has the effect of making it possible to engage in work without being aware of the table of contents during the writing process. .

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a diagram showing an example of the operation of the device shown in Fig. 1, and Fig. 3 is a diagram showing an example of the operation of the device shown in Fig. 1. item·
A diagram to explain which subcategory it is classified into,
FIG. 4 is a diagram for explaining the operation of a conventional table of contents generation device. 1: Original text input section, 2: Table of contents item candidate extraction section, 3: Morphological analysis section, 4: Sentence final word analysis section, 5: Table of contents classification section, 6: Table of contents generation result display section. E 1 Figure

Claims (1)

[Claims] (1) An automatic table of contents generation device that automatically extracts table of contents items in a text converted into an electronic text file, classifies them into large, medium, and small table of contents items, and automatically creates a table of contents for the text. It consists of an original text input section, a table of contents item candidate extraction section, a morphological analysis section, a sentence final word analysis section, a table of contents classification section, and a table of contents generation result display section, and the original text input section accepts input of the digitized text file, In the above table of contents item candidate word extraction section, the number of characters in each character string from the beginning of a sentence to the next line break, from a line break to the next line break, or from a line break to the end of a sentence in the electronic text file from the original text input section is within a predetermined value. Extracts character strings whose first character is an Arabic numeral, Roman numeral, Chinese numeral, or an Arabic numeral, Roman numeral, Chinese numeral enclosed in parentheses such as a bullet or lock, or a symbol. The morphological analysis unit performs morphological analysis on the character string extracted by the table of contents item candidate extraction unit, and the sentence-final word analysis unit determines whether the sentence-final word of the character string subjected to morphological analysis by the morphological analysis unit is a noun. The above-mentioned table of contents classification section extracts a character string that is a suffix or a suffix, and determines whether the first character of the sentence is an Arabic numeral, Roman numeral, Chinese numeral, round brackets, or square brackets for the character string extracted by the sentence-final word analysis section. Surrounded Arabic numerals, Roman numerals,
Classified into large items, medium items, and small items based on whether they are Chinese numerals or symbols, and in the table of contents generation result display section above,
An automatic table of contents generation device characterized in that the character strings classified by the table of contents classification section are displayed on a display device as a table of contents.