JP2638777B2 - Kana-Kanji conversion device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a kana-kanji conversion device that converts a kana character string input to input a Japanese sentence into a sentence mixed with kana-kanji characters.

[Prior Art] Conventionally, in a kana-kanji conversion device having an independent word dictionary that stores only the stem of a word, learning of words, phrases, and kana-kanji mixed sentences has been roughly divided into the following methods.

(1) Rewrite the use order information only for the existing stem part of the independent word dictionary.

(2) Words and phrases are stored in the main memory as separate dictionaries in the main storage device, and at the time of Kana-Kanji conversion, the contents of the dictionary for learning are given priority over the contents of the independent word dictionary.

(3) A dictionary is stored in the auxiliary storage device in units of words and phrases as a separate dictionary, and at the time of kana-kanji conversion, the contents of the dictionary for learning are given priority over the contents of the independent word dictionary.

However, in the method (1), if “long” is selected from the phrase candidates such as [Nagai-long-long-unnamed], “long” is registered in the independent dictionary as only “long”. But
Since “Nagai” is registered as “Nagai”, “Long” can be output in preference to “Long”, but there is a drawback that “Nagai” cannot be output in priority.

In addition, although the method (2) does not have the drawbacks of the method (1), since a learning dictionary is created in the main storage device, the main storage device requires a large area, and the area for operating other systems becomes large. In addition to the decrease, the area for this can be relatively small, and there is a disadvantage that the learning content is lost due to the power-off in the case of the device whose contents are lost when the main storage device is turned off.

Although the method (3) solves the problems (1) and (2), the double search of another dictionary takes a long processing time and the number of stored character strings increases explosively. was there.

[Problems to be Solved by the Invention] As described above, if the ending of a word is registered in the independent word dictionary, the storage content is increased. Therefore, it is basically desired to store only the stem of the word. If learning is performed in units, it is impossible to learn in units of phrases. For this reason, if dictionaries for learning are created in the main storage device and the auxiliary storage device, the use area increases and the storage content increases. Processing speed is delayed.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above problems and to provide a user-friendly device without increasing the capacity of a dictionary.

[Means for Solving the Problems] The present invention relates to a kana-kanji conversion device for converting an input kana character string into a kanji-kana mixed sentence while selecting a homophonetic word, comprising: An independent word dictionary having a kanji part corresponding to the heading part, a character string candidate determining means for securing a desired kanji kana mixed candidate character string, and the character string candidate determining means for obtaining a desired kanji kana mixed candidate character string. With the determination, the longest length of the heading part consisting of the stem of the independent word present in each candidate character string other than the determined candidate character string is calculated as the longest independent word heading length. Comparing the longest independent word heading length with a fixed independent word heading length that is the length of a heading portion of the determined candidate character string, the heading being composed of the stem of the independent word; Independent word watching Longer than the length,
A learning character string creating means for creating a part corresponding to the heading of the longest independent word heading length in the determined candidate character string as a learning character string; and the created learning character strings in the highest priority selection order. And an independent word dictionary rewriting means for registering in the independent word dictionary.

[Operation] In the present invention, when the longest independent word heading length is longer than the determined independent word heading length, a portion corresponding to the heading portion of the longest independent word heading length in the determined candidate character string is created as a learning character string. .

[Embodiment] Here, details of the present invention will be described based on an illustrated embodiment.

FIG. 1 is a diagram showing a kana-kanji conversion device according to one embodiment of the present invention. In the figure, reference numeral 1 denotes a central processing unit (hereinafter abbreviated as CPU) for controlling the entire system based on data input from the input unit 2. The character string read by the input unit 2 is sent to the kana-kanji conversion device 7, where the kana-kanji conversion is performed.

In the kana-kanji conversion device 7, the kana character string captured by the kana-character input unit 8 is converted by the grammatical analysis unit 9 into a sentence mixed with kana-kanji while searching the attached word dictionary 10 and the independent word dictionary 11, The candidates are stored in the candidate storage unit 12 as a set of homophonic phrase candidate groups in units of phrases. At this time, regarding the independent words, taking into account the use order information stored in the independent word dictionary, the inside of the homophone phrase candidate group is arranged in order from the longest independent word heading. Those having the same length are arranged in the order of use of the independent words.

When the phrase candidate selection unit 3 requests a phrase candidate from the phrase candidate output unit 13, the phrase candidate output unit 13 sends the phrase candidate selection unit 3 in accordance with the order of the candidate phrases stored in the phrase candidate storage unit 12. Outputs phrase candidates. The phrase candidate selection unit 3
The phrase candidates are displayed one after another on the phrase candidate display unit 4, and the user determines the phrase to be obtained by the phrase candidate determination unit 5 while watching the phrase candidates.

The information of the phrase determined by the phrase candidate determination unit 5 is passed to the determined phrase reception unit 14. Here, the learning word creation unit 15
Retrieves homonymous phrase candidates of the target phrase stored in the phrase candidate storage unit 12 based on this information, and finds the longest independent word heading length of the longest phrase in the independent word stem portion from the search results. The headword length is compared with the length of the independent word heading of the determined phrase. If the length is equal to or longer than the longest independent word heading length, learning of the homonym word in the independent word dictionary (that is, the order of use) is performed. And the independent word rewriting unit 16 searches the independent word dictionary 11 for the independent word stem part of the definite phrase, and deletes the learning word in the homophonic word group if it exists. At the same time, the use order of the homophone word group is rewritten so that the use order of the independent word stem portion of the definite phrase becomes first, and the independent word dictionary is updated.

When the longest independent word heading length is longer than the independent word heading length of the confirmed phrase, the learning word creation unit 15 learns the kana-kanji mixed portion of the fixed phrase corresponding to the longest independent word heading length of the confirmed phrase. Created as a word, add a symbol that can be identified as a learning word and the part of speech of the final word in the original composition of the learning word composed of a plurality of words, make it into a form that can be registered in the independent word dictionary, and hand over. The independent word rewriting unit 16 searches the independent word dictionary 11 for a homophone word group corresponding to the heading of the learning word, and if a learning word already exists in the homophone word group, deletes the learning word and deletes it. , The use order of the homophone word group is rewritten so that the use order of the new learning words becomes first, and the independent word dictionary is updated.

By these operations, the independent word dictionary can have the learning words and the ordinary independent word stems in the order of use, and the inside of the homophone / syllable clause candidate group of the kana-kanji conversion word firstly contains the heading of the independent word. Despite the fact that they are learned by arranging them in order from the longest one, the independent word headings are short, eliminating the disadvantage of not being output preferentially.In the next kana-kanji conversion, conversions that reflect the learning results are possible Becomes

FIG. 2 is a flowchart for explaining the operation of one embodiment of the present invention. The operation will be described below with reference to FIG.

Consider a case where “long” is input as a kana character string. At this time, the independent word dictionary 11 is configured as shown in FIG.

Here, in <200>, if the selected candidate is “long”, the independent word dictionary 11 has an independent word stem of “long”, so the independent part heading length is two digits. Next, in <201>, the longest independent part heading length is searched. In this case, there is “Nagai” as a homophone. Since “Nagai” is registered in the independent word dictionary 11, the longest independent word heading length is three digits.

Next, in <202>, the heading length (two digits) of the selected candidate and the longest heading length (three digits) are compared, and the heading length of the candidate for which the longest heading length is selected is selected. Since it is longer, the learning word of <203> is created.

That is, the word “long” corresponding to the heading (long) corresponding to the longest head of the independent part is created as a learning word. In this case, the word “long” is assigned a code that is a learning word (not shown). Next, in <204>, the part of speech of the learning word “long” is created from the part of speech of the last word.

Next, in <205>, it is checked whether or not there is an existing learning word in the independent word dictionary. And if not,
The learning word is registered in the independent word dictionary so that the learning word has the highest priority. If there is an existing learning word, the existing word is deleted and “long” is registered in the independent word dictionary in place of the existing learning word. (<206>) That is, in the present embodiment, the independent word portion (“Nagai”) of a candidate phrase other than the selected phrase is read from the independent word portion (“Naga”) of the selected phrase. If the is longer, it is assumed that up to the longest reading length (long) of the independent word portion of the phrase in the candidate other than the selected phrase is the independent word portion of the selected phrase, and the independent word + adjunct (Or conjugation ending)
In the unit (“long”), the part of speech is used as the part of speech of the last word, and a symbol indicating that it is a learning word is added, and the word is newly stored in the independent word dictionary. When storing the learning word, if there is an existing learning word, the existing learning word is deleted and only the new learning word is stored.

Although not described in the description of the flowchart of FIG. 2, the independent word portion of the candidate phrase other than the selected phrase has a shorter or equal reading than the independent word portion of the selected phrase. In this case, the highest priority use order information is added to the content of the independent word dictionary corresponding to the independent word stem portion of the selected phrase.

 Further, the details will be described below.

When "Nagai" is selected in the above embodiment, only the use order information of "Nagai" in the independent word dictionary is given top priority, and the storage amount of the dictionary does not increase.

If "Long" is selected, "Long" is stored as a learning word, and this use order information is stored in the independent word dictionary with the highest priority, and the highest priority is output by the next input.

When "long" is selected, "long" is stored as a learning word. However, since "long" is already stored, this is deleted and "long" is stored. Therefore, an increase in the storage amount of the dictionary can be minimized.

[Most-bloomed-torn-cracked], [congratulatory-glue and-
The same is true for the examples such as "
Since the appearance of such a candidate group is relatively small in Japanese, the storage amount of the dictionary is practically small.

FIG. 3 is an explanatory diagram showing the operation of the above embodiment. In this example, when the input character string "Nagai" is washed first with the word "Long" and then again with "Long" 2 shows an internal structure of the independent word dictionary 11 of FIG. 11a is a heading part, 11b is a kanji part, and 11c is a priority part.

Initially, the priority word dictionary 11 is configured as shown in FIG. 3 (a). At this time, the display priority order of “Nagai” is “Nagai”, “Unknown”, “Long”, “Long” (“Nagai”). Is the highest priority. The same applies hereinafter). In this state, "Long" will not be displayed in preference to "Nagai".

Next, when "long" is selected, the character "long" is inserted among the three kanji characters, and an identifier (not shown) indicating that this character string is a learning word is added. The configuration is as shown in FIG. At this time, the display priorities for “long” are “long”, “long stay”, “unknown”, and “long”.

Further, this time, when "long" is selected for "long", "long" is replaced with "long" as shown in FIG. 3 (c), and the identifier of the learning word is changed as before. Attached. The display priorities in this rotation are “long”, “long stay”, “unknown”, and “long”. Where "long"
Is skipped because "Long" is displayed first.

[Effects of the Invention] As described above in detail, the present invention provides that when the longest independent word heading length is longer than the determined independent word heading length, the heading portion of the longest independent word heading length in the determined candidate character string is Since there is a learning character string creating means for creating a corresponding part as a learning character string, the conversion efficiency can be significantly improved while minimizing the increase in the capacity of the dictionary.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a kana-kanji conversion apparatus according to one embodiment of the present invention. FIG. 2 is a flowchart showing the operation of the kana-kanji conversion apparatus according to one embodiment of the present invention. FIG. 3 is a diagram describing an example of a change in the content of an independent word dictionary according to the operation of the kana-kanji conversion device according to one embodiment of the present invention. 7 ... Kana-Kanji conversion device 11 ... Independent dictionary

Claims (1)

(57) [Claims] 1. A kana-kanji conversion device for converting an input kana character string into a sentence mixed with kanji or kana while selecting a homophone word, comprising a heading portion composed of a stem of an independent word and a kanji portion corresponding to the heading portion. An independent word dictionary, a character string candidate determining means for determining a desired Kanji kana mixed candidate character string, and the character string candidate determining means determining a desired Kanji kana mixed candidate character string, Among the lengths of the heading part consisting of the stem of the independent word present in each candidate character string other than the candidate character string, the longest length is calculated as the longest independent word heading length, and the longest independent word heading length, Comparing with the determined independent word heading length, which is the length of the heading part consisting of the stem of the independent word in the determined candidate character string, and when the longest independent word heading length is longer than the determined independent word heading length, Confirmed Learning string creating means for creating a portion corresponding to the heading portion of the longest independence word heading length in the candidate character string as a learning character string; and A kana-kanji conversion device, comprising: an independent word dictionary rewriting means for registering in a dictionary.