JPH04305B2 - - Google Patents

Description

[Detailed description of the invention]

[Technical Field of the Invention] The present invention relates to a document creation device that can efficiently create Japanese sentences without placing a large burden on an operator. [Technical background of the invention and its problems] In recent years, Japanese sentences have been inputted in units of phrases as kana character strings, which are phonetic information, using kana keyboard devices, and independent word parts in the phrases have been converted into kanji. Document creation devices that create Japanese documents have become widespread. However, when creating Japanese documents using this device, errors in inputting kana characters become a major problem. Generally, errors in inputting kana characters tend to occur more often at the end of conjugated words. The reality is that in conventional document creation devices, the stem of the word converted into Kanji and the conjugated ending of the kana character that caused the error are displayed as they are on the screen of the display device. Detection of such errors is left to the operator to monitor the displayed document. For this reason, detecting errors requires the operator to watch the display screen, placing a heavy burden on the operator, and correcting discovered errors is also complicated. Moreover, as described above, since the detection of errors is left to the operator, errors are often overlooked. This has resulted in a substantial decline in document creation efficiency. [Object of the Invention] The present invention has been made in consideration of the above circumstances, and its purpose is to effectively detect and correct errors in kana character input without placing a burden on the operator. An object of the present invention is to provide a highly practical document creation device capable of efficiently creating Japanese documents. [Summary of the Invention] The present invention finds candidate words by comparing a kana character string input in a bunsetsu unit with an independent word dictionary, and when the candidate word is a predicate, the candidate word ( The conjugated endings are tested for the character strings excluding the candidate words and the conjugated endings, and the character strings excluding the candidate words and the conjugated endings from the kana character strings are checked against an adjunct dictionary. The connection condition with an independent word (word) is checked, and if the connection condition fails, the conjugated ending is changed to generate a correct kana character string candidate and this is output on the monitor. . [Effect of the Invention] Thus, according to the present invention, depending on whether the connection condition between an independent word and an attached word in a kana character string constituting one clause is met, the kana ending of the conjugated word when the independent word is a term is determined. Errors in character input can be effectively detected. Then, a correct kana character string candidate is generated by changing the conjugated ending that caused the error, and this kana character string candidate is output on the monitor, so the operator can obtain this information and easily input correction information. becomes possible. Therefore, it is possible to easily and efficiently create a Japanese document without placing a large burden on the operator. Furthermore, by outputting the kana character string candidates to the monitor using synthesized speech, the effort of keeping a close eye on the display screen can be significantly reduced, thereby making it possible to prevent operator fatigue. In addition, it is possible to efficiently correct errors in inputting kana characters at the endings of frequently used words, and to speed up document creation. [Embodiment of the Invention] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The figure is a schematic configuration diagram of an embodiment of the apparatus, in which kana character strings indicating the phonetic sequence of Japanese sentences are inputted in units of phrases via the keyboard device 1. Note that a voice input device may be used in place of the keyboard device 1, and Japanese sentences may be input by voice. In this case, the input speech may be cut into phrases and a phoneme sequence may be formed for each phrase. Thus, the input control circuit 2 temporarily stores the kana character string input in units of phrases as described above, and first supplies this to the independent word matching circuit 3. This independent word matching circuit 3 matches the above-mentioned kana character string with headwords described later in an independent word dictionary registered in advance in the independent word dictionary storage device 4, and selects independent words that can be matched as candidate words. It is stored in the word stack storage device 5. The independent word dictionary registered in the dictionary storage device 4 consists of headwords of many independent words in kana notation, grammatical information, information in kanji notation, accent information, and information on appearance frequency. Then, the matching circuit 3 performs matching by detecting a match between the headword of each independent word and the input kana character string, and stores the matched independent words as candidate words in the storage device 5 in descending order of the number of characters. It has become a staple. Now, when the candidate word stored in the candidate word stack storage device 5 is a predicate, that information is given to the inflection ending test circuit 6. Whether or not the independent word is a term is informed by the grammatical information of the independent word dictionary. The test circuit 6 uses the conjugation ending table storage device 7 to test the conjugation endings of the remaining character strings from the input kana character strings after removing the candidate words (original stems). This test of conjugated endings is performed in order from the first ranked candidate word until a passing result is obtained in the test. When the candidate words and their inflected endings that have passed the test are obtained, the adjunct word test is performed by the adjunct word test circuit 8. This adjunct word test is
This is performed using the adjunct word dictionary storage device 9 for the remaining character strings after removing the candidate words and their conjugated endings from the input kana character string. The attached word dictionary is
The attached word and the pre-conjunctive conditions under which the attached word can be connected,
In other words, it is given as the part of speech, conjugation type, conjugation form, etc. of the independent word that can be connected to the attached word, and the verification circuit 8
will test whether these connection conditions are met. The following table shows an example of an adjunct word dictionary.

【table】

Claims (1)

[Scope of Claims] 1. Means for comparing a kana character string input in a bunsetsu unit with an independent word dictionary, means for storing independent words matched by this means as candidate words, and means for storing independent words that are matched by this means as candidate words. At some point, means for testing the conjugation ending of a character string obtained by removing the candidate word (original language stem) from the kana character string, and a character string obtained by removing the tested conjugation ending and the candidate word from the kana character string. A method for checking the attached word with an attached word dictionary and testing the conjunctive condition of the attached word with respect to the independent word (phrase), and changing the conjugation ending of the kana character string that failed the test of this conjunctive condition to correct it. A document creation device comprising: means for generating kana character string candidates; and means for monitoring and outputting the kana character string candidates generated by the means. 2. The document according to claim 1, wherein the means for monitoring and outputting kana character string candidates generates voice parameters of the generated kana character string candidates and outputs synthesized speech according to the voice parameters. Creation device.