JPH05342260A - Word spelling check device - Google Patents

Abstract

PURPOSE:To display candidate words corresponding to an inputted word with parallel translation at the time of checking the erroneous spelling of the word inputted to a translation device and to narrow the candidates down by the part of speech at the time of deciding the part of speech. CONSTITUTION:An unknown word detection means 12 detects an unknown word from a sentence to be translated. A syntax analytic means 11 decides the part of speech of the unknown word by a part of speech decision means 13. A candidate word and parallel translation selection means 14 selects the candidate words corresponding to the unknown word to display on a display part 6. When the part of speech decision means 13 decides the part of speech, the candidate words can be narrowed down.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improvement of an apparatus for checking spelling of words in a sentence to be translated in a translator for translating a foreign language into Japanese.

[0002]

2. Description of the Related Art A description will be given of the case of an English-to-Japanese translator which is now widely used in Japan.

FIG. 9 is a block diagram of a conventional English-Japanese translator.

This is an input section A for inputting a sentence to be translated, a storage section B for temporarily storing the input sentence,
A translation unit C that performs translation processing. The input unit A is a device for inputting English data to be processed by the translation unit C, and includes a keyboard 1, a floppy disk drive 2, a character recognition device 3 having a scanner, and the like. The storage section B has a text data file 4 for storing the English text data input by the input section. The translation unit C is centered on a translation processing device 5 that performs translation processing on the data stored in the text data file 4, a grammar rule for translation and a dictionary 7, a display unit 6 that displays translation contents at any time, It is composed of a translated text data file 9 for storing translated translated data, a printer 8 for outputting a translation result, and the like. When a sentence to be translated is input, it is translated by the translation processing device 5, and finally the translated sentence is output by the printer 8.

[0005] In the conventional method of checking an English sentence input by a translator as described above, the grammar rule and the grammar rule are checked by matching the character string of the input word with the grammar rule and the character string of the word registered in the dictionary. A word of a character string that is not in the dictionary is extracted. This word is hereinafter referred to as an unknown word. Then, candidates for words close to the character string of the unknown word are displayed. This unknown word is considered a misspelling of the word.

[0006]

As described above, in the past, only candidate words similar to an unknown word are displayed, and the entire sentence is not referred to. Therefore, no related part of speech candidate is also displayed. Was there. In addition, since the dictionary including the translated word is not referred to, the candidate result is displayed only in English words, and there is a case where it is difficult to make a judgment because there is no selection criterion.

[0007]

In the word spelling check device of the present invention, the syntax analysis means for the sentence to be translated, the unknown word detecting means in the sentence to be translated, and the sentence to be translated are translated. A part of speech determination unit for an unknown word detected when the syntax is possible, a candidate word and a parallel translation selecting unit corresponding to the unknown word, and a display unit for these are provided.

[0008]

[Function] The structure of each word is examined by the syntactic analysis means,
Check if the syntax is translatable. Then, if there is an unknown word in the translatable syntax, the part of speech is limited by the relationship before and after in the syntax, and candidates for the same part of speech near the character string can be displayed, and the representative translated word can also be displayed. ..

Even when the part-of-speech is not determined, it is possible to display close candidates of the character string of the word and their parallel translations.

[0010]

FIG. 1 is a block diagram of an embodiment of the present invention. This can be realized by adding necessary means to the translation unit C in FIG.

Foreign texts such as English texts from the text data file 4 are sent to the syntax analysis means 11 inside the translation processing device 5. Here, the syntax is analyzed for each sentence. The syntactic analysis is performed by the part of speech of the words in the sentence and their arrangement. Then, the translated sentence is displayed on the display unit 6 for each translatable sentence.

When creating a translated sentence, the syntactic analysis means 11 refers to the grammar rule and the dictionary 7, and if there is a word in the translated sentence that is not in the grammar rule and the dictionary 7, that is, an unknown word, the unknown word detection means 12 determines the word. Is displayed on the display unit 6 with a predetermined identification (for example, diagonal lines).

In the translated text displayed on the display unit 6, there are words that do not make sense because of the context. This is the spelling in the dictionary, but not the word intended by the original author. The same spelling may have other meanings. Also at this time, the suspicious word is given a predetermined identification. In these cases, the user determines.

In the translated text displayed on the display unit 6, there is a word that is not in the dictionary or a word whose meaning is unknown, but the part of speech may be determined based on the context. In that case, the part-of-speech determination means 13 determines the part-of-speech for the suspicious word.

Words that are not in the dictionary can be automatically detected by the unknown word detecting means 12. In addition, for example, the word following the numeral can be automatically determined as a noun by the part-of-speech determining unit 13. This facilitates narrowing down candidate words.

Next, referring to the grammatical rules and the dictionary 7, words near the character string of the unknown word are selected by the post-candidate and parallel translation selecting means 14 and displayed on the display unit 6.

FIG. 2 is a flow chart showing the concrete processing contents of the spelling check, and FIGS. 3 and 6 show the first and second sentences to be translated for explaining the operation of the processing contents. This is a concrete example. 4 and 5 are views showing the results after the English sentence shown in FIG. 3 is processed by the spell check, and FIGS. 7 and 8 are subjected to the respective processes corresponding to FIG. It is the figure which displayed the result after.

First, a word error checking process for the sentence shown in FIG. 3 will be described. This is an example of a sentence in which a word in the sentence that should be “are” is “aer”.

In step S1, the translation processor is 1
Translate sentences. In the translation process, first, the part of speech of each word is checked using a dictionary. Next, based on the grammar rules, the syntactic analysis means 11 determines whether or not the sentence is formed by those parts of speech.

In the sentence of FIG. 3, since there is no verb,
The sentence does not hold and the translation fails (step S2).
Since the translation is not successful, the flow proceeds to step S3. In step S3, a dictionary is used to check whether there is an unknown word. When an unknown word is found, the flow proceeds to step S4, and candidates as shown in FIG. 4 are displayed. Candidate words with similar character strings are selected from the dictionary and displayed together with the representative translation. Based on the displayed result, the user determines whether or not there is a correct candidate (step S5), selects the correct candidate (step S6), and if not, inputs the correct word (step S7).

In the case of FIG. 4, since "are are" is a correct word, when "are" is selected, the correction result is corrected as shown in FIG. 5, and the corrected translation result of the English sentence is displayed.

Next, a word spelling check process for the sentence of FIG. 6 will be described. The word in the sentence is "bo
An example where “books” is a place that should be “oks”.

In step S1, translation is performed in the same manner as described above.

Since the necessary elements are present in the structure of the sentence shown in FIG. 6, the translation will be successful (step S
2) Then, in step S8, it is determined whether or not there is an unknown word. In the case of the sentence in FIG. 6, there is “books”, but this is after “ten”, and in step S9, the part of speech that “books” is a noun is determined.

Next, the flow advances to step S4.

As a result, the candidates as shown in FIG. 7 are displayed together with the bilingual translation. In addition to selecting from the words having a close character string,
Since the part of speech is determined to be a noun, candidates are narrowed down. Based on the displayed result, the user performs the same operation as that described with reference to FIG. 3 (steps S5 to S
In 7), the translation result of the corrected English sentence as shown in FIG. 8 can be obtained.

It is also possible to first detect an unknown word and then perform syntax analysis.

[0028]

As described above, according to the present invention, since the selection candidates are displayed with the translated words in the dictionary and the candidates are narrowed down to the part of speech determined by applying the translation process, the efficient words are selected. Spell check can be realized. Also,
Since you can see the full translation result, you can check the whole sentence.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is an example of a flowchart of the present invention.

FIG. 3 is an example of a sentence having a misspelled word.

FIG. 4 is an example of a candidate corresponding to a misspelled word.

FIG. 5 is an example of a sentence in which an error is corrected.

FIG. 6 is another example of a sentence having misspelled words.

FIG. 7 is an example of a candidate corresponding to a misspelled word.

FIG. 8 is an example of a sentence in which an error is corrected.

FIG. 9 is a block diagram of a conventional translator.

[Explanation of symbols]

 4 Text Data File 6 Display 7 Grammar Rule and Dictionary 11 Parsing Means 12 Unknown Word Detecting Means 13 Part of Speech Determining Means 14 After Candidate and Parallel Translation Selecting Means

Claims (1)

[Claims] 1. A syntactic analysis unit for a sentence to be translated, an unknown word detecting unit in the sentence to be translated, and a part of speech of an unknown word detected when the sentence to be translated has a translatable syntax. A word spelling check device comprising: a determining unit, a candidate word corresponding to an unknown word and a parallel translation selecting unit, and a displaying unit.