JPH0283770A - Sentence sort deciding device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Application Field) The present invention relates to devices such as word processors and computers that create, edit, and proofread texts, and in particular, relates to devices such as word processors and computers that create, edit, and proofread texts. The present invention relates to a text classification/judgment device having a function of analyzing and classifying sentence types, comparing and examining the sentence types of consecutive sentences, and detecting errors or possible mistakes.

(Prior art) When creating and editing sentences using devices such as word processors and computers, it is not enough to simply create and edit, but also to check whether the created or edited sentences are free from errors or to make them persuasive. It would be very desirable if you could check that the text is correct.

When creating a text for the purpose of conveying one's opinions and thoughts to others, it is desirable that the text be persuasive.

In order to make your writing persuasive, you need to be creative, such as not indicating that the opinions and ideas in your writing are based on facts. Therefore, before or after a sentence that expresses an opinion or idea, put a sentence that expresses the facts that serve as the reason. On the other hand, if a sentence is a series of sentences without expressing an opinion, or if the sentences before and after the sentence expressing an opinion are hearsay from a third party, the opinion and the facts on which it is based can be separated. The cause-and-effect relationship becomes unclear, the writing lacks persuasive power, and the purpose of conveying opinions and ideas cannot be fully achieved.

When proofreading your written work, you must pay attention to the following points. First, each sentence in the text to be proofread is classified as either the author's opinion, hearsay from a third party, or fact. Next, we conduct an investigation to determine whether there are any consecutive sentences expressing an opinion, and whether the parts before and after the opinion are hearsay.

Here, sentences ending with "I think...", "I think...", etc. are considered to be "opinions."Also,"~
It seems. ” can be classified as hearsay. These sentences can be easily distinguished by whether or not a specific word such as ``think'', ``think'', or ``dasoda'' is present at the end of the sentence.

After classifying the sentence types in this way, the above-mentioned judgment can be made by comparing the sentence types of consecutive sentences with several patterns to support grammar proofreading.

(Problem to be solved by the invention) However, there is no conventional word processor or computer that has the function of classifying sentences into sentence types and supporting proofreading, etc. Ta. Therefore, judgments and calibrations based on these classifications are
Humans had to watch everything and do it. Therefore, after creating a text, it takes a lot of effort to make it persuasive and easy to understand, and proofreading also takes time. In addition, oversights and errors are likely to occur, creating a heavy burden.

In view of such conventional circumstances, the present invention classifies each sentence in a text to be proofread into sentence types such as opinion, hearsay, and fact, and determines the sentence types of consecutive sentences to detect errors or mistakes. It is an object of the present invention to provide a text classification/judgment device having a function of detecting portions where there is a possibility of .

[Structure of the Invention] (Means for Solving the Problem) The text classification and determination device according to the present invention includes a classification means for classifying each sentence in the target text into text types such as fact, hearsay, and opinion;
The apparatus also includes a determination means for comparing the sentence types of each sentence classified by the classification means and detecting an error or a possible error as a sentence type of consecutive sentences.

(Operation) With a text classification/judgment device having such a configuration, for each sentence in the target text, the classification means first identifies the type of sentence, such as fact, hearsay, opinion, etc., based on the presence or absence of a specific word in the sentence. are categorized. Based on the sentence type of each sentence classified in this way,
Their continuation is compared and examined as a pattern by the determining means, and errors or possible errors are detected.

In this way, sentences are judged based on the sentence type, and places where there is a possibility of errors are detected in terms of how the sentences continue, and proofreading of sentences, which previously relied entirely on manual labor, is now possible. You will be supported to be able to do it more accurately in a shorter time.

(Embodiment) FIG. 1 is a block diagram showing the main parts of a grammar proofing support device according to an embodiment of the present invention.

There is an input unit 1 comprising, for example, a keyboard, for inputting character information and various instructions. There is also a display section 8, which is made up of, for example, a display, and which displays the text under proofreading and highlights the locations where errors or possible errors occur (hereinafter referred to as warning locations).

In addition, in this example, the text to be proofread is based on facts, hearsay,
In order to classify sentences into types such as opinions, each sentence is first broken down into morphemes, which are the individual words that make up each sentence, and analyzed.

The morphological analysis unit 2 performs this morphological analysis, and uses the analysis dictionary 3 to analyze sentences while breaking them down into individual words.

A classification unit 4 is provided for classifying the sentence type for each sentence decomposed into morphemes. Classification rule 5 is
This file is a file containing, for example, a table, the classification criteria used when the classification unit 4 classifies sentence types.

FIG. 2 shows an example of a table included in this classification rule 5. According to the rules shown in this figure, each sentence is classified into sentence types such as "opinion" and "hearsay."

The sentence types of each sentence classified in this way are extracted as patterns according to the continuation of the sentence, and compared and examined with Judgment Rule 7 to detect errors or possible error (warning) parts in the sentence. There is part 6. Judgment rule 7 is
For example, it is a file that has a table of patterns of sentence types that are erroneous or have the possibility of errors.

FIG. 3 shows an example of this determination rule 7. For example, the sentence types of three consecutive sentences are "opinion", "opinion", and "opinion".
If the pattern is lined up with , the table is set so that it is determined as an error that "opinions continue."

FIG. 4 is a flowchart 1 showing an example of the operation of this embodiment device. In this flowchart, sentences are classified and judged each time for sentences set in sequence.
This shows an example of the operation in which a warning point is displayed if there is one. Hereinafter, an example of the operation of this apparatus will be explained based on the flowchart of FIG.

First, according to instructions from the human resources department 1, sentences to be classified and judged are set. This sentence may be a part of a document input in advance, or it can be set to be classified and determined while being input sequentially (steps 101 to 1-03 in FIG. 4).

FIG. 8 shows an example of sentences set as targets for classification and determination in this way. ``I work.'' ``I think it's right.'' ``I think he's suspicious.''
I don't think she understands. The targets are sentences consisting of a series of sentences such as ``.

The morphological analysis unit 2 decomposes each sentence into individual morphemes based on the morphological analysis dictionary information held in the analysis dictionary 3, and the morphological analysis results A list of is generated (step 105).

FIG. 5 shows an example of a list of example sentences that have been morphologically analyzed. Information such as each morpheme that makes up a sentence, its pronunciation, and part of speech is displayed as a list.

Based on the list I of this morphological analysis result, the classification unit 4
Accordingly, each sentence is classified into sentence types based on the classification rule 5. Here, the unit of a sentence is from the beginning of the sentence or the previous period to the end of the sentence or the next period. The classified results are set in the classification result list (step] 07).

FIG. 6 is an example of a classification result list showing the sentence type of each sentence created in this way. The first sentence of the sentence shown as an example in FIG. 8 is classified as "fact", and the second to fourth sentences are all classified as "opinion".

Next, based on the determination unit 6 or the determination rule 7, a warning location is determined and detected from this classification result list. This determination result is held as warning data (step 109).

FIG. 7 is a diagram showing an example of warning data determined in this manner. In the example sentence in Figure 8, since the second to fourth sentences are all "opinions," it is determined that they fall under the error "Opinions continue." according to the determination rule shown in Figure 3. Detected as a warning point.

As a result of classifying and determining the target text in this way, if a warning part that is retained as warning data is detected (step 1-1 affirmative), for example, that part is highlighted on the display screen of the display unit 8. and corresponding warning data are displayed (step 113).

In this way, by simply setting a sentence to be classified and judged, each sentence in that sentence is classified into a sentence type, and is judged as a sentence type of consecutive sentences, thereby eliminating errors or potential errors. A certain location will be detected.

[Effects of the Invention] As explained above in detail, the text classification/judgment device according to the present invention classifies each sentence of the target text into text types such as fact, hearsay, and opinion, and distinguishes between consecutive sentences. By comparing, examining, and determining each item individually, errors or potential errors can be detected without the need for human intervention.

Therefore, for example, when proofreading text, etc., it is possible to significantly reduce the time required for proofreading, improve accuracy, greatly reduce the burden on proofreaders, etc., and improve understanding. It becomes possible to obtain easy-to-understand sentences.

Furthermore, if the text classification/judgment device according to the present invention is used when reading a text created by another person, it will be of great help in understanding the text, and will also help determine whether the author's opinions and ideas are valid. It will also help you decide.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a grammar proofing support device according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of a classification rule in the same embodiment, and FIG. 3 is a diagram showing an example of a judgment rule in the same embodiment. 4 is a flowchart showing an example of the operation of the apparatus of the embodiment, and FIG. 5 is a diagram showing an example of a morphologically analyzed list.
Figure 6 shows an example of a list of classification results into sentence types, Figure 7
The figure is a diagram showing an example of determined warning data, and FIG. 8 is a diagram showing an example of a sentence set as a target for classification and determination. 1... Input section 2... Morphological analysis section 3... Analysis dictionary 4... Classification section 5... Classification rule 6... Judgment section 7... Judgment rule display section

Claims (1)

[Claims] (1) A classification means that classifies each sentence in the target text into sentence types such as fact, hearsay, opinion, etc., and compares the sentence types of each sentence classified by this classification means and determines that the sentence types of consecutive sentences are incorrect. A text classification/judgment device characterized by comprising: a determination means for detecting a location where there is a possibility of an error.