JPS5918730B2 - Kana/Kanji conversion device - Google Patents

Description

[Detailed Description of the Invention] The present invention utilizes a kana/kanji conversion device, particularly the so-called longest match method, and grammatical information, which is information indicating connection relationships between preceding and following words, to perform kana/kanji conversion and input. In a kana/kanji conversion device that inserts a space to indicate a break between words in data, if the next word following the space does not have a connection relation in terms of grammatical information with the word before the space, the corresponding grammatical The present invention relates to a kana/kanji conversion device that discards a short message and selects the next word.

Generally, in a kana/kanji conversion device that inputs a kana sentence and converts it into a sentence mixed with kanji, the words stored in the word dictionary and the reading of the input kana that is converted have the largest number of characters. The longest match method, which selects the word that matches as the correct word, is widely used.

On the other hand, check the grammatical connection between the previous word selected and the next word to be selected,
The next word is selected on the condition that the correct connection relationship is obtained. Generally, words are converted as described above, but in order to make the above conversion process more efficient, when the operator inputs words, a space is also given to indicate a break between words.

If this blank space is given, it is assumed that one word ends at that blank position. This method of inserting spaces gives better results than filling in spaces, but on the other hand, the grammatical information determined just before the space is used to select the next word following the space. If this is an erroneous conversion, it will have an adverse effect on the conversion of the word immediately after the blank, and there is a possibility that it will be erroneously converted one after another. This is because it is considered absolutely correct that one word ends with the space, and the previous word selected to satisfy this condition is considered correct. The present invention aims to solve this problem, and the kana/kanji conversion device of the present invention includes a word dictionary corresponding to readings and a grammar dictionary storing grammatical conditions for words that follow at least the selected word. and a dictionary search device for indexing the word dictionary and the grammar dictionary using the blanks as delimiters for input data input with readings and with spaces between the input words. The kana/kanji conversion device converts the above input data into a sentence containing kanji based on the longest match method and the grammatical information for the previous selected word. A grammatical information check unit is provided to check grammatical information between the next word and the grammatical information check unit detects that the grammatical information cannot satisfy a predetermined condition for the next word. In this case, the dictionary search device is configured to instruct the dictionary search device to discard the grammatical information and select the next word by assuming that the next word is a word following a period or comma of the sentence. The dictionary search device is characterized in that it is configured to perform a search corresponding to the instruction.

This will be explained below with reference to the drawings. FIGS. 1A, B, and C are explanatory diagrams for explaining the prerequisite problems of the present invention, FIG. 2 is an embodiment showing the storage contents of a word dictionary, and FIG. 3 is an embodiment showing the storage contents of the grammar dictionary. FIG. 4 shows the configuration of an embodiment of the present invention.

It is now assumed that a kana character string 2 as shown in FIG. 1A is input in order to obtain a character string 1 as shown in FIG. 1C. In this case, spaces 3 are inserted between words, etc., in order to efficiently perform the conversion process. However, when the kana sentence 2 shown in FIG. 1A is input and the conversion process is performed, the conversion process may become impossible at some point, as in character string 4 shown in FIG. 1B. This means that the word "da" (illustrated A) before the blank 3 is the final form "da" in the word dictionary 5 illustrated in Figure 2.
This is due to the fact that it was selected. Such selection does not occur based on the priority based on the frequency of occurrence of the final form "da" and the adjunctive form "da", and in the case of the method of inserting blank 3, this selected final form This arises from the fact that ``da'' is easily regarded as absolute due to the existence of a blank space. When the final form "da" is selected in this way, the result of the selection is such that, from the grammatical information in the grammar dictionary 6, nouns cannot be connected as the next word (× in Figure 3). mark), and the next word "system" cannot be connected, and the unconvertible state continues.In addition, in the above case, narrow down the heading of the word "da" in the word dictionary to one. The grammatical information for "da" can be either the final form or the adnominal form, so for example, three pieces of information are written in the form of "final form", [adjunctive form], and [adjunctive/final form]. Just create a word dictionary.

However, when the word dictionary 5 is created in this way, the grammar information storage area to be stored in the word dictionary 5 becomes large, and the word dictionary becomes large because an area of the same size must be prepared in advance for all words. I don't like it. Also, when converting, it is essential to correctly select one of the three pieces of grammatical information listed above. In some cases, the possibility of erroneous conversion increases compared to the method in which . For this reason, in the case of the present invention, if the next word "system" becomes unconvertible as a result of selecting the final form [da] as described above, the grammatical information of the said "da", that is, the final form. To discard information to the effect that there is a certain point and to process it as if the period or comma in the sentence was the previous word.

That is, the grammatical information of a period or a comma is read from the grammar dictionary 6, the connection relationship with the next word "system" is checked, and the word "system" is converted. In the illustrated grammar dictionary 6 in FIG. 3, a noun can be connected to a comma (indicated by a circle in the figure), and the word "system" is correctly converted, as in character string 1 shown in FIG. 1C. In addition, the above-mentioned situation where conversion becomes impossible is as follows.
It is not limited to the above-mentioned final form and continuous form,
For example, this may occur in the following cases. In other words, “beautiful moving
This also occurs when a dictionary is used that does not allow an adjective (in this case, ``beautiful'') to be connected with the adnominal form of a verb (in this case, ``moves''), such as when inputting ``horse''. . FIG. 44 shows the configuration of an embodiment of the present invention, where numerals 5 and 6 correspond to FIGS. 2 and 3, and 7
1 is a character string input device (for inputting the character string 2 shown in FIG. 1A); 8 is a buffer register in which, for example, the character string up to the blank 3 is set and subjected to conversion processing; 9 10 is a dictionary section, 10 is a dictionary search device, 11 is a period or comma grammar information reading address section, 12 is a grammar information check section, and 13 is an output file.

If the character string ``Kumikonda'' is now set in the register 8, the dictionary search device 10 will search for the characters ``Yuo Ku'', ``Kumi'',
The word dictionary 5 is indexed in order for each of "Kumiko", "Kumikon", and "Kumikonda", and "Kumiko (incorporated)" is selected based on the longest match method and transferred to the output file 13 side.

In this case, assuming that the connection to the previous word "o" is correct from the perspective of grammatical information, the grammatical information check section 1
2 sends the 0K signal to the output file 13 and performs the above "incorporation"
is determined. Next, it is assumed that the dictionary search device 10 determines "n" and also determines the next "da" as the final form "da".
Next, the word dictionary 5 is indexed for the character string "SYSTEM" set in the buffer register 8, and the grammar dictionary 6 is indexed based on the final form obtained from the previously determined [da]. In this case, the string "system" is 1
However, since the word is a noun, the grammar information check unit 12 returns a NO signal to the dictionary search device 10. The dictionary search device 10 reads the character string “SYSTE”
, [sis], "shi", etc., but due to the grammatical information and the presence of a blank space after "mu", the correct conversion cannot be obtained. In this case, in the present invention, the grammar information (final form) for [da] is discarded, and the grammar dictionary 6 is accessed based on the contents of the address 11,
Learn that nouns can be connected as the next word. As a result, the next word is determined to be "system" and is transferred to the output file 13, and the grammar information check circuit 12 gives a 0K signal to the output file 13, so that the word "system" is selected. As explained above, according to the present invention, while enjoying the advantages of the method of inserting spaces and without increasing the capacity of the word dictionary, subsequent erroneous conversions caused by incorrect selection of the word before the space can be avoided. It becomes possible to prevent this.

[Brief explanation of drawings]

Figures 1A, B, and C are explanatory diagrams explaining the prerequisite problems of the present invention, Figure 2 is an example showing the memory contents of a word dictionary, and Figure 3
The figure shows an example of the storage contents of a grammar dictionary, and FIG. 4 shows the configuration of an example of the present invention. In the figure, 1, 2, 4 are character strings, 3 is a blank, 5 is a word dictionary,
6 is a grammar dictionary, 7 is a character string input device, 8 is a buffer register, 9 is a dictionary section, 10 is a dictionary search device, 12 is a grammar information check section, and 13 is an output file.

Claims (1)

[Claims] 1 A word dictionary corresponding to the reading and a grammar dictionary storing grammatical conditions for at least the words following the selected word are provided, and the word dictionary is input with the reading and a space is provided between the input words. A dictionary search device is provided that indexes the word dictionary and the grammar dictionary using the blank space as a delimiter, and searches the input data using the longest match method and grammatical information for the previous selected word. A kana/kanji conversion device that converts a sentence mixed with kanji based on the above-mentioned blank space is provided with a grammar information check unit that checks the grammar information between the word selected as a delimiter and the next word that follows, and the grammar information When the checking unit detects that the grammatical information does not satisfy a predetermined condition for the next word, the checking unit discards the grammatical information and instructs the dictionary search device to discard the grammatical information so that the next word becomes The dictionary search device is configured to issue an instruction to select the next word by assuming that the word follows a period or comma in a sentence, and the dictionary search device is configured to perform a search corresponding to the instruction. Kana/kanji conversion device.