JPH07282046A - Kana-kanji conversion method - Google Patents

Abstract

PURPOSE:To automatically register a certain amount of words by connecting continuous noun groups, registering them as one word and setting the reading of an initial noun constituting the noun group to be the reading of the registered word. CONSTITUTION:When an inputted sentence is key-inputted like 'Toranomon Minato Ward, Tokyo', for example, a KANA (Jpanese syllabary)-KANJI (chines character) conversion dictionary 1 divides the inputted sentence into words 2-l to 2-6 which can be converted by the KANA-KANJI conversion dictionary 1 itself. After conversion by the KANA-KANJI conversion dictionary 1 is decided, the noun is extracted from the converted sentence and whether the plural nouns continuousy exist in the sentence or not is judged. Such a noun group is set to be one word and it is automatically registered. A registration control part 4 executes the registration processing, and 'Toranomon Minato Ward, Tokyo' is registered as one new word in such a case. The part of speech is the noun and the reading of 'Tokyo' being the initial word 2-1 in the noun sentence is adopted for the reading.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a kana-kanji conversion method used for inputting Japanese into a word processor or other information processing device, and more particularly to a kana-kanji conversion method for automating the word registration and conversion.

[0002]

2. Description of the Related Art A word processor and various information processing apparatuses are provided with a kana-kanji conversion system for converting input kana into kanji in order to facilitate Japanese input. When using such a system, word registration is used so that frequently used words can be easily input. Generally, when a word is registered using the Kana-Kanji system, the registered word, the reading of the registered word, and the part of speech of the registered word are input. by this,
For example, when an address "Toranomon, Minato-ku, Tokyo" is input, if "Tokyo" is input and the conversion key is pressed, the sentence "Toranomon, Minato-ku, Tokyo" is converted.

[0003]

By the way, in order to perform the word registration as described above, it is generally necessary to call a subsystem dedicated to the word registration and perform the registration procedure using the subsystem. In this case, the user must input the reading and the part of speech. However, it is very cumbersome to register a word that is expected to be frequently used during text entry by calling the word registration sub-system each time, which causes a delay in the text creation work.

[0004]

In order to solve the above problems, the present invention automatically registers words as follows. First, when a plurality of nouns are continuously converted into corresponding kanji while inputting a sentence using the kana-kanji conversion dictionary, these continuous noun groups are connected to form one word. Then, the reading of the first word constituting the noun group is set to the reading of one concatenated word and the words are automatically registered. The dictionary registered in this way is called a complementary dictionary.

When the character of the first noun is input when converting the input sentence into kanji, it is checked whether or not there is another noun reading before and after the reading of the first noun. Then, when the reading of another noun is not involved, the Kanji conversion by the complementary dictionary is executed, and in other cases, the Kana conversion by the general Kana-Kanji conversion dictionary is executed.

[0006]

In this method, when a sentence is input, it is determined whether or not a plurality of nouns are continuously present in the sentence. When a plurality of nouns are continuously converted into corresponding kanji, a group of continuous nouns is concatenated into one word. For that reading, the reading of the first noun that constitutes the noun group is adopted. As a result, while inputting a sentence, words with consecutive nouns are automatically registered, facilitating subsequent conversion input.

[0007]

The present invention will be described in detail below with reference to the embodiments shown in the drawings. FIG. 1 is an explanatory view showing an embodiment of an automatic word registration method of the present invention. This figure shows a kana-kanji conversion dictionary 1 for converting an input sentence into kana-kanji and a registration control unit 4 for executing word registration according to the present invention to generate a complementary dictionary 3. That is, as shown in this figure, for example, when an input sentence is key-in such as "Tokyo and Minato Tokuranonon", the Kana-Kanji conversion dictionary 1 can convert these by the Kana-Kanji conversion dictionary 1 itself. Word 2-1
~ 2-6. In addition, in the Kana-Kanji conversion dictionary 1,
For example, the following information is stored and the conversion operation is performed.

FIG. 2 is a diagram for explaining a kana-kanji conversion operation.
As shown in the figure, the kana-kanji conversion dictionary 1 stores, for example, reading of words, converted words, and their parts of speech in association with each other. In this example, "Tokyo" is a kanji converted from "Tokyo" and its part of speech is a noun. Dictionary information is similarly set for other words. By referring to such dictionary information, the sentence after Kanji conversion for the input sentence shown in FIG. 1 becomes “Toranomon no, Minato-ku, Tokyo”. Such a conversion result is displayed on a display or the like, and the input person looks at it and presses the enter key.

In the present invention, after the conversion by the Kana-Kanji conversion dictionary 1 is confirmed, the word registration is automatically executed as follows. First, a noun is extracted from the converted sentence. Then, it is determined whether or not a plurality of nouns continuously exist in this sentence. In general, when two or more nouns are used consecutively, such words are often used again by being similarly connected. In the present invention, focusing on such a point, such a noun group is automatically registered as one word. This registration processing is performed by the registration control unit 4 in the figure. That is, in this example, "Toranomon, Minato-ku, Tokyo" is registered as a new word. The part of speech is a noun. In addition, the reading adopts the reading "Tokyo" about "Tokyo" which is the first word of the noun sentence.

That is, as shown in this figure, "Toranomon, Minato-ku, Tokyo" is newly registered in the complementary dictionary 3. Therefore,
After that, if you input "Tokyo", it will be converted into the registered character "Toranomon, Minato-ku, Tokyo".

FIG. 3 shows a more specific processing operation flowchart of the method of the present invention. First, when a sentence is input in step S1, Kanji conversion is performed for each word constituting the sentence in step S2, and the confirmation process is executed. Then, in step S3, the part of speech of each word is searched and cut out. Then, in step S4, it is determined whether the word is a noun. If it is a noun, the process returns to step S3 to check the part of speech of the next word.

Here, in step S4 again, it is judged whether or not it is a noun, and if it is not a noun, the process proceeds to step S5. In step S5, it is determined whether or not the processing loop of steps S3 and S4 is repeated twice or more,
Determine whether two or more nouns are consecutive. If two or more nouns are consecutive, the automatic word registration of the present invention is performed. In this case, the process proceeds to step S6, and information of a word to be newly registered is generated from the word cut out as a noun.
That is, a new word is obtained by connecting two or more cut-out nouns. Next, in step S7, it is determined whether the new word needs to be registered.

The determination as to whether or not this registration is necessary is to avoid duplicate registration, which will be described later. If it is determined that registration is necessary, the process proceeds to step S8, and the newly generated word is registered in the complementary dictionary.

FIG. 4 shows an apparatus block diagram for carrying out the method of the present invention. When carrying out the method of the present invention, specifically, for example, an apparatus configured as shown in this figure is used. First, a kana-kanji conversion dictionary 1 is provided for converting kana-kanji of an input sentence. Further, a word decomposition rule unit 5, a word decomposition unit 6, a word part-of-speech search unit 7, etc. for analyzing and processing sentences are provided. Further, in order to register a word in the complementary dictionary 3, a registration portion cutout rule unit 8, a registration portion cutout unit 9, a registration information generation rule unit 10, a cutout word information generation unit 11, and a registration necessity rule unit 12 are registered. , Registration necessity determination unit 1
3, the cut-out word registration unit 14 and the like are provided. Also,
For referring to the complementary dictionary, a complementary dictionary reference rule unit 15, a complementary dictionary reference determination unit 16, a complementary dictionary search unit 17, a registered word replacement unit 18, and the like are provided.

The word decomposing unit 6 divides the sentence, which has been input and has undergone the conversion of Chinese characters, into words constituting the sentence. The word part-of-speech search unit 7 is a part that searches the part-of-speech by searching the Kana-Kanji conversion dictionary 1. Further, a registration part cutout rule part 8 and a registration part cutout part 9
When a noun is continuous according to the procedure described in FIG. 3, is a part for performing a process of sequentially cutting out the continuous noun. The registration information generation rule unit 10 and the cut-out word information generation unit 11 are parts that collectively put the cut-out noun groups into words to be registered. The registration necessity rule part 12 and the registration necessity judgment part 13 are parts for making a registration judgment described later, and the cut word registration part 14 is the first part of the noun group as already described with reference to FIG. This is a part for registering the reading of the noun of as the reading of the word. Each of these parts is composed of each module of a program that constitutes the Kana-Kanji conversion system.

On the other hand, the complementary dictionary reference rule unit 15 and the complementary dictionary reference determining unit 16 determine whether to refer to the Kana-Kanji conversion dictionary 1 or the complementary dictionary 3 when a reading to be described later is input. It is a part. Complementary dictionary search unit 17
Is a part for extracting a registered word from the complementary dictionary 3 when it is determined to search the complementary dictionary 3. The registered word replacement unit 18 is a part that replaces the reading part of the input sentence with the word registered in this way. These are also composed of program modules that make up the Kana-Kanji conversion system.

FIG. 5 shows an explanatory diagram of the automatic word conversion method of the present invention. When words are automatically registered as described above, if a sentence that should be converted by referring to the Kana-Kanji conversion dictionary as usual is converted by unnecessarily referring to the complementary dictionary, incorrect conversion will occur. easy. Therefore, in the case of implementing the present invention, whether or not to refer to the complementary dictionary is determined by the following method.

First, as a rule, when the reading of a word registered in the complementary dictionary is input and no other noun is included before and after the reading, the complementary dictionary is referred to.
That is, for example, as shown by * 1 in FIG. 5, when "Tokyo" is input, "Tokyo" is a noun and "no" is a particle. Therefore, the noun "Tokyo" is entered alone. At this time, referring to the complementary dictionary 3, the conversion “of Toranomon of Minato-ku, Tokyo” is performed.

On the other hand, as shown by * 2, when there is an input such as "Tokyo Metro", "Tokyo" and "eki" are both nouns. In this way, when another noun “Eki” exists along with the reading “Tokyo” registered in the complementary dictionary, the Kana-Kanji conversion dictionary 1 is referred to and a normal conversion is executed. As a result, as shown in this example, the normal conversion "at Tokyo station" is executed. In this way, erroneous conversions such as "at Toranomon Station, Minato-ku, Tokyo" can be prevented. As shown in the example of * 3, when the single word "Tokyo" is input, the complementary dictionary 3 is also referred to.

Next, in the present invention, it is preferable to adopt the following method in order to prevent the processing that the already registered words are erased or redundantly registered in the automatic registration. . FIG. 6 shows an explanatory diagram (part 1) of the update registration method. In this example, using the procedure described above,
The case where the word "Tokyo" is already registered with the reading "Tokyo" will be described. In this case,
It is assumed that the word "Minato-ku, Tokyo" is newly generated and its reading is nominated as "Tokyo". In the present invention, when it is determined that the same reading is already registered in the complementary dictionary, the old word and the new word are compared. Such a comparison determination process is performed by the complementary dictionary reference determination unit 16 which has already been described with reference to FIG.

At this time, it is determined whether the new word includes the old word. In this example, the new word "Minato-ku, Tokyo" includes the old word "Tokyo". Moreover, the new word is longer than the old word. Therefore, in this case, the old word is replaced with the new word and registered with the same reading. In this way, from the next input, the same reading of "Tokyo" can be used to convert the longer character "Minato-ku, Tokyo".

FIG. 7 is an explanatory diagram (part 2) of the update registration method. In this example, it is assumed that the registered old word is read as "Tokyo" and is registered as "Toranomon, Minato-ku, Tokyo". In this case, the new words newly registered as candidates for registration have the same reading, “Minato Ward, Tokyo”. Here, comparing the old word with the new word, the old word is not included in the new word, and the old word is longer than the new word. Therefore, if a new word is replaced by an old word and registered, the words that were registered in the long noun group will be replaced by short words,
The effect of converting all at once becomes thin.

Therefore, in such a case, the registration necessity judgment unit 13 shown in FIG. 4 prohibits the registration in that case so that the old word is not replaced with the new word. Note that the new registration process is useless even when the old word and the new word are exactly the same. Therefore, registration is prohibited even in such a case. In this way, by replacing the old noun group registered with the same reading with the new new noun group, a longer word is automatically registered and the conversion efficiency is improved.

The present invention is not limited to the above embodiments. In the above embodiment, the place name is shown as an example of the registration target, but various words such as company name, person's name, article name, etc. can be automatically registered for words obtained by connecting a plurality of nouns. .

[0025]

The kana-kanji conversion method of the present invention described above concatenates consecutive noun groups when a plurality of nouns are continuously converted into corresponding kanji while inputting a sentence using the kana-kanji conversion dictionary. Then, it was registered as one word, and the reading of the first noun that composes the noun group was set to the reading of the registered word. Therefore, it is not necessary to call a general word registration dedicated subsystem to perform registration work. Also, it becomes possible to automatically register a certain word. Therefore, the user does not have to be aware of word registration, and the conversion efficiency of Chinese characters can be further improved. Moreover, the user does not need to input the reading or the part of speech of the word to be registered, which reduces the burden on the user.

When the reading of a word registered in the complementary dictionary is input, if there is no reading of another noun before or after the reading, the kanji conversion by the complementary dictionary is executed. In other cases, the normal kana-kanji is used. By executing the Kanji conversion by the conversion dictionary, it is possible to prevent unnecessary reference to the complementary dictionary and prevent erroneous conversion. In addition, erroneous conversion is reduced because the complementary dictionary does not convert words composed of a plurality of nouns that the Kana-Kanji system does not know.

Furthermore, the old noun group that has already been registered is compared with the new noun group that is about to be newly registered, and the new noun group is included only when the new noun group includes the old noun group and is longer than the old noun group. If the registration is performed, the longest word is always registered in the complementary dictionary. Therefore, the contents of the complementary dictionary can be replaced so as to improve the conversion efficiency. Moreover, by preventing duplicate registration, the complementary dictionary can be efficiently maintained and managed with a small capacity.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of an automatic word registration method of the present invention.

FIG. 2 is an explanatory diagram of a kana-kanji conversion operation.

FIG. 3 is a flow chart of the method of the present invention.

FIG. 4 is an apparatus block diagram for performing the method of the present invention.

FIG. 5 is an explanatory diagram of an automatic word conversion method of the present invention.

FIG. 6 is an explanatory diagram (1) of an update registration method.

FIG. 7 is an explanatory diagram (part 2) of the update registration method.

[Explanation of symbols]

 1 Kana-Kanji conversion dictionary 3 Complementary dictionary 4 Registration control unit

Claims (3)

[Claims] 1. When a plurality of nouns are continuously converted into corresponding kanji while inputting a sentence using a kana-kanji conversion dictionary, these consecutive noun groups are connected and registered as one word, A kana-kanji conversion method characterized by setting the reading of the first noun that constitutes a noun group to the reading of a registered word. 2. A kana-kanji conversion dictionary for converting a basic word in an input sentence into kanji, as well as a continuous noun group that is registered as one word and is the first noun constituting the noun group. Is set to the reading of the registered word, a complementary dictionary is provided to convert the registered word into kanji, and when the reading of the first noun is input, it is different before and after the reading of the first noun. A kana-kanji conversion method, wherein kanji-kanji conversion is performed by the complementary dictionary when the reading of the noun is not performed, and kanji-kanji conversion is performed by the kana-kanji conversion dictionary in other cases. 3. When a plurality of nouns are continuously converted into corresponding kanji while inputting a sentence using a kana-kanji conversion dictionary, these consecutive noun groups are connected and registered as one word, When setting the reading of the first noun that composes a noun group to the reading of a registered word, if a word with the same reading has already been registered, the old noun group that has already been registered will be newly registered. A kana-kanji conversion method characterized in that the new noun group is compared and the new noun group is registered only when the new noun group includes the old noun group and is longer than the old noun group.