JPS62150461A - Sentence error detection system - Google Patents

Abstract

PURPOSE:To contrive efficient proofreading of a Japanese sentence by using a concordant word dictionary storing sets of adverbs and auxiliary verbs having a concordance so as to check the error in the usage of words having a correspondence. CONSTITUTION:A concordant adverb retrieval means 3 retrieves a concordant adverb stored in the concordance relation work dictionary 1 from a Japanese sentence stored in a sentence storage means 2 and the detected concordant adverb and the position of it in the sentence are outputted to a corresponding auxiliary verb retrieval means 4 and an error suggestion means 5. The retrieval means 4 references the dictionary 1 to obtain an auxiliary verb to the concordant adverb detected and the auxiliary verb corresponding to the adverb is retrieved from a character string between the concordant adverb and a sentence end symbol closest to the adverb and after it. When the auxiliary verb corresponding to the adverb is not detected, the error suggestion means 5 indicates the concordant adverb as an error in the Japanese sentence.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a method for detecting usage errors in input Japanese sentences.

(Prior Art) In recent years, Japanese word processors have become widespread, and the efficiency of creating Japanese sentences has improved. However, efficiency has improved in aspects such as inputting, saving, and reusing Japanese texts, and it cannot be said that efficiency has improved in terms of proofreading, which still relies on human hands.

(Problem to be solved by the invention) In Japanese sentences, for example, as shown below, there are sentences in which adverbs and sentence-final expressions (auxiliary verbs) correspond.

It probably doesn't seem like there's any more...but in reality, Sometimes we end up creating sentences that do not follow this consonance relationship. For example, the following sentence:

, "Maybe it will rain tomorrow."
However, the expression at the end of the sentence is in the form of an assertion, which is unnatural in Japanese. Correctly, this sentence should be rewritten as follows, using the inferential auxiliary verb ``u'' at the end of the sentence.

``It's probably going to rain tomorrow.'' Detecting such errors in pragmatics has traditionally required humans to actually read the text and check it, but it has been hoped that automatic checking by machines would be possible. Ta.

An object of the present invention is to provide a method for automatically checking the usage of words that have consonant relationships such as those described above, which conventionally relied on acquisition when proofreading Japanese texts.

(Means for Solving the Problems) Means provided by the present invention to solve the above-mentioned problems is a sentence error detection method that detects errors in input Japanese sentences, which have a consonance relationship. A consonant relational word dictionary that stores pairs of adverbs and auxiliary verbs, a consonant adverb search means for searching the input Japanese text for adverbs stored in the consonant relational word dictionary, and adverbs detected by the consonant adverb search means. A corresponding auxiliary verb search means searches for an auxiliary verb having a consonance relationship with the character string between the adverb in the input Japanese sentence and the sentence final symbol closest to the end of the adverb; The present invention is characterized by comprising an error presentation means for presenting the detected adverb as an error when a related auxiliary verb is not detected.

(Example) The structure and operation of the present invention will be explained in detail using the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the text error detection method of the present invention. Each component will be explained.

In FIG. 1, reference numeral 1 denotes a consonant relational word dictionary, which stores pairs of adverbs and auxiliary verbs having a consonant relation.

FIG. 2 shows an example of a vocative related word dictionary. In FIG. 2, lO is an adverb of response, and 11 is a corresponding auxiliary verb.

In FIG. 1, reference numeral 2 denotes a text storage means, which stores input Japanese texts. It can be realized using an IC memory, a magnetic disk device, a magnetic tape device, etc.

Reference numeral 3 denotes a consonant adverb retrieval means, which searches for consonant adverbs stored in the consonant related word dictionary 1 from Japanese sentences stored in the sentence storage means 2. This can be achieved by repeating matching of the Japanese sentence and the 10 parts of the consonant relational vocabulary 1411. When a consonant adverb is detected in the Japanese sentence, the corresponding auxiliary verb search means 4 and the error presentation means 5
Tohe, outputs the detected consonant adverb and its position in the sentence.

Reference numeral 4 denotes a corresponding auxiliary verb search means.
Obtain the auxiliary verb that corresponds to the auxiliary adverb by referring to the auxiliary verb, and then find the auxiliary verb that corresponds to the auxiliary adverb from the character string between the auxiliary adverb and the closest sentence-final symbol after the auxiliary adverb. Search for. This can be achieved by repeatedly matching the character string with the 11 parts of the consonant relational word dictionary 1. Then, it outputs to the error presentation means 5 whether or not an auxiliary verb corresponding to the adverb of response has been detected.

Reference numeral 5 denotes an error presentation means which, when the corresponding auxiliary verb search means 4 does not detect an auxiliary verb corresponding to a choric adverb, displays the position of the choric adverb sent from the choric adverb searching means 3. Presenting vocative adverbs as errors in Japanese sentences. As a presentation method, there is a method of showing the position of the corresponding adverb in the target Japanese sentence by plinking, reversing, underlining, etc. on the display.

Next, the operation of this embodiment will be explained using an example.

It is assumed that the sentence inputted and stored in the sentence storage means 2 is as follows.

``It's probably going to rain tomorrow. It's like the rainy season.'' Let us also assume that the vocative relational word dictionary 1 has the contents as shown in Figure 2.

First, the consonant adverb search means 3 matches the sentence with 10 in the consonant relational word dictionary 1, and detects "maybe" at the beginning of the sentence. Then, the detected "maybe" and its position (first to third character) are matched to the auxiliary verb search means 4 and the error presentation means 5.
Output to. In response, the corresponding auxiliary verb search means 4 refers to the auxiliary relational word dictionary 1, obtains ``u'' as an auxiliary verb corresponding to ``maybe'', and combines this ``u'' with ``maybe'' in the sentence and its successor. Search from between the nearest sentence ending symbol ".", that is, from "It's going to rain tomorrow."

As a result, "U" was not detected, and the incorrect presentation of that fact was made.
Output to. In response to this, the error presentation means 5 treats the consonant relation of "maybe" as an error and changes the position of "maybe" to a blank/
Presented by means such as

Furthermore, the same process is repeated immediately after "maybe".

The consonant adverb search means 3 detects ``as if'' from the 10th to the 12th character, and the corresponding auxiliary verb search means 4 detects ``as if'' as the corresponding adverb for ``as if''.
is detected from the following character string ``It looks like the rainy season.'' As a result, the error presentation means 5 does not present "Toyorude" as an error.

In the end, in the target sentence, ``maybe'' is presented as having an error in the vocative relation.

In the example shown in FIG. 1, the detection of adverbs of consonance and the corresponding auxiliary verbs were performed by matching character strings.
By performing matching on the morphological analysis results, it is possible to perform matching more accurately.

For example, in the case of the following sentence, in the example of Figure 1,
By simply matching character strings, the ``u'' at the end of the conjugated verb is incorrectly identified as the auxiliary verb ``u'' which corresponds to ``maybe'', and is determined not to be an error.

``Maybe I'll buy a book.'' In this case, if we perform morphological analysis, we will get the following sequence of words, which can be determined to be incorrect because the auxiliary verb ``u'' that corresponds to ``light-fun'' is missing.

Maybe (adverb) buy a book (noun) (case particle) (verb stem)
U (verb conjugation ending) FIG. 3 is a block diagram showing the configuration of an embodiment of the present invention in the case of matching with the morphological analysis results.

In FIG. 3, l is a consonant relational word dictionary, 2 is a sentence storage means, 3 is a consonant adverb search means, 4 is a corresponding auxiliary verb search means, 5 is an error presentation means, 6 is a morphological analysis means, 7 is a word dictionary, 8 is an analysis result storage means. The morphological analysis means 6 detects words from the sentences stored in the sentence storage means 2 while referring to the word dictionary 7, performs a grammatical connection test, and writes the sentences into the analysis result storage means 8 as a word series. The morphological analysis means 6, the word dictionary 7, and the analysis result storage means 8 are known techniques. The search for consonant adverbs in the consonant adverb search means 3 and the search for corresponding auxiliary verbs in the corresponding auxiliary verb search means 4 are performed by matching the series of words stored in the analysis result storage means 8 with the words in the consonant relation word dictionary 1. However, the operation of the embodiment is almost the same as that in FIG.

(Effects of the Invention) As explained above, according to the present invention, it is possible to automatically detect parts of a created sentence where there is a possibility that there is an error in the usage of words in a consonance relationship. It is possible to improve the efficiency of proofreading work for Japanese texts, which previously relied on manual labor.

[Brief explanation of drawings]

FIGS. 1 and 3 are block diagrams showing the configuration of an embodiment of the present invention, and FIG. 2 is a diagram showing an example of a vocative relational word dictionary. 1... Consonance related word dictionary, 2... Sentence storage means, 3.
...Method for searching for consonant adverbs, 4.Means for searching for corresponding auxiliary verbs,
5... Error presentation means, 6... Morphological analysis means, 7.
...Word dictionary, 8...Analysis result storage means.

Claims (1)

[Claims] In a sentence error detection method for detecting errors in input Japanese sentences, a vocative relational word dictionary that stores pairs of adverbs and auxiliary verbs having a vocative relation, and adverbs stored in the vocative relational word dictionary are used in the input Japanese sentences. A consonant adverb search means to search from a Japanese sentence, an auxiliary verb that has a consonance relationship with an adverb detected by the consonant adverb search means, and a auxiliary verb in a consonant relationship with the adverb in the input Japanese sentence and the end of the sentence that is closest after the adverb. A corresponding auxiliary verb search means for searching from a character string between a symbol and an error presenting means for presenting the detected adverb as an error when the corresponding auxiliary verb search means does not detect the auxiliary verb in the consonance relation. A text error detection method characterized by: