JPH0785238B2 - Dictionary search device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a dictionary search device in a kana-kanji conversion device, and more particularly to a dictionary search device that appropriately converts an input character string including a long vowel symbol. is there.

2. Description of the Related Art In recent years, sentence creation devices such as Japanese word processors have become popular. This type of device generally inputs a kana character into a created sentence and converts it into a character sequence containing a mixture of kanji characters for each word to create a Japanese sentence.

By the way, recently, a method of input processing Japanese information by inputting voice and mode information designating a character type corresponding to the voice has been proposed, but many problems in inputting a language by voice are Has occurred.

In particular, when a character string including a long vowel symbol is input, how the kana character is converted into the long vowel symbol to search the word dictionary greatly affects the search time. For example, if the input voice contains the character string "e", even if you search the word dictionary with the kana character string that has converted the long vowel symbol into the kana character "e", such a kana character string will appear in the dictionary. In many cases, it was necessary to search again with another kana character string that was newly converted to another kana character “i”.

The present inventors have previously proposed Japanese Patent Application No. 59-9959 as a dictionary search device.
The one shown in No. 1 was proposed.

FIG. 3 shows a block diagram of this dictionary search device. 1 is a character string containing a long vowel symbol, and the long vowel symbol next to "Adan" is "A" or "Idan". The next long vowel symbol and "i", the next long vowel symbol next to "Udan" is "U",
The long vowel symbol next to "Edan" is "I" or "E",
A character string conversion unit that converts the long vowel symbol next to "Odan" to "U" or "O" and outputs it when the long vowel symbol is not included. The designated dictionary 3 is a dictionary collation unit for searching the dictionary 2 with the character string obtained from the character string conversion unit 1.

The operation of the dictionary search device configured as described above will be described below. For explanation purposes, Express with.

For example, suppose that when the user wants to input “closed”, the recognition result when the syllables “he” and “sa” are input by voice is as shown in Table 1.

At this time, input character string It is assumed that the candidates shown in Table 2 are obtained as the candidate of the character string for.

When this character string candidate is input to the character string conversion unit 1, the character string is converted according to the rules in Table 3.

Table 4 shows the result of character string conversion according to the rules of Table 3.

Then, the character string candidates shown in Table 4 are sent to the dictionary matching unit 3. Since Japanese writing words are registered in the dictionary 2, the dictionary collating unit 3 collates with the dictionary in order from the first candidate of the character string, and outputs the corresponding dictionary content as a correct answer. In the case of this example, the character string "closed" to be obtained can be obtained only after searching the dictionary 13 times.

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention However, in the above configuration, when a character string having a long vowel symbol next to "Edan" as shown in the example is input, the long vowel symbol is uniformly distributed. In addition, since the character strings converted into "i" and "e" are collated with the dictionary, there is a problem that it takes a lot of time to collate the dictionary.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a dictionary search device capable of conveniently and easily processing long vowels in Japanese speech information.

Means for Solving the Problems In the present invention, when a character string including a long vowel symbol is input, the syllable preceding the long vowel symbol is focused, and the front syllable is “Adan”.
When is, the long vowel symbol is "A", when it is "Idan", it is "I",
"Udan" is converted to "U", "Edan" is converted to "I" or "E", and "Odan" is converted to "U" or "O" for output. When the syllable that precedes the long vowel is "Edan" when searching the dictionary with the character string conversion unit that outputs the string as is, and the character string obtained from the character string conversion unit Is a dictionary search device including a dictionary collating unit that searches a dictionary with a character string in which a long vowel symbol is uniquely converted into “i” or “e” according to the type of a pre-syllable.

With the above-described configuration, the present invention converts a long vowel symbol into a character string including a long vowel symbol subsequent to the "edan" syllable and converts the long vowel symbol into a character, and when searching the dictionary, , By limiting the character string to be searched, an appropriate Kana-Kanji conversion output can be efficiently obtained.

First Embodiment FIG. 1 shows a block diagram of a dictionary search device according to the present invention. In FIG. 1, 1 is a character string converter, 2 is a dictionary,
The above is the same as the configuration of FIG. 3 is a character string obtained from the character string conversion unit 1, and when the dictionary 2 is searched, if the syllable preceding the long vowel symbol is "Edan", the long vowel symbol is selected according to the type of the front syllable. It is a dictionary matching unit that searches a dictionary with a character string that is uniquely converted to "i" or "e".

The operation of the dictionary search device of this embodiment configured as described above will be described below. For explanation purposes, Express with.

For comparison, description will be made using the character strings described in FIG. When you want to input the character string "closed", when the recognition result when you input the monosyllabic "he" and "sa" as shown in Table 1, the input character string The candidates shown in Table 2 are obtained as the character string candidates for the. When this character string candidate is input to the character string conversion unit 1, the character string is converted according to the rules of Table 3, and the character string candidates as shown in Table 4 are sent to the dictionary matching unit 3. The dictionary collation unit 3 searches the dictionary according to the rules of Table 5 when searching the dictionary for character string candidates as shown in Table 4.

According to the rule of Table 5, among the character string candidates shown in Table 4,
It is only necessary to search the dictionary for the character string candidates shown in Table 6.

In this case, the character string "closed" can be obtained by the seventh dictionary search, and it is understood that almost half of the character strings to be obtained are obtained by the dictionary search as compared with the conventional example.

As described above, according to this embodiment, the dictionary search time can be significantly reduced by using the conversion rule for long vowel symbols in Table 5 described above.

The conversion rules for long vowel symbols shown in Table 5 are based on the bias of existence as Japanese words as shown below.

First, a word containing a continuous vowel was extracted from 30,000 Japanese words. Here, a continuous vowel refers to a vowel in which two or more vowels are continuous in one word or in word concatenation. In this case, each vowel syllable is not limited to a sequence of independent vowel syllables, but a sequence of a vowel included in the preceding syllable and a vowel syllable immediately after it is also referred to as a continuous vowel.

For example, "love", "nephew", "fish", etc. are continuous vowels, and "kai", "carp", "sui". Etc. also contain continuous vowels. When there are continuous vowels, some vowels are isolated from each other, and some vowels are closely linked to the immediately preceding vowel, and one vowel sounds like a long vowel. The latter one will be referred to as a "heavy vowel".

As a result of investigating a word including a continuous vowel in the 30,000 Japanese words in the dictionary 2 used in this device, the following data were obtained.

There are only 10 words that contain the character string "Edan" + "E". Of these, the words that can be uttered in the long sound as "Echo" are "Akanbee", "Ehto", "Heysan", and "Heyaya".
4 words.

There are 1520 words that include the character string “Edan” + “I”, of which 1495 is the word that can be uttered in the long sound.
There is a word. And, in this, there is no word including the character string "nei".

The dictionary 2 used in this device has the above-mentioned bias rules and the present invention has a conversion rule for long vowels based on this rule.

The rules of the present embodiment are based on the bias of the dictionary used in this device, and it is natural that the bias will be slightly different if the dictionary used is different.
Therefore, the present invention has rules according to the dictionary used,
The feature is that the character string to be searched is limited before the dictionary search.

Further, in the present embodiment, the dictionary search character string is limited by the dictionary matching unit, but the same effect can be obtained at any stage before the dictionary matching. For example, when limiting by the character string converting unit, , The limitation may be made by converting the long vowel symbols into only the character string to be actually searched.

EFFECTS OF THE INVENTION As described above, according to the present invention, a character string including a long vowel symbol is searched by a character string obtained by converting the long vowel symbol into a character according to the type of the syllable. Since it is configured to search the dictionary by limiting the character string,
An appropriate Kana-Kanji conversion output can be efficiently obtained even for a character string including a long vowel symbol, and its practical effect is great.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram showing a configuration of a dictionary search device according to an embodiment of the present invention, FIG. 2 is an operation explanatory diagram of the same embodiment, and FIG. 3 is a schematic configuration diagram of the dictionary search device previously proposed. is there. 1 ... Character string conversion unit, 2 ... Dictionary, 3 ... Dictionary collation unit.

Claims (1)

[Claims] 1. When a character string including a long vowel symbol is input, paying attention to a syllable that precedes the long vowel symbol, and when the preceding syllable is "Adan", the long vowel symbol is "A". , "Idan" indicates the long vowel symbol "i", "Udan" indicates the long vowel symbol "U", and "Edan" indicates the long vowel symbol "I" or ""E","Odan", the long vowel symbol is converted to "U" or "O" and output, and when the long vowel symbol is not included, the character string conversion unit and the dictionary are output. A dictionary matching unit that searches the dictionary with a character string obtained from the character string converting unit, wherein the dictionary matching unit determines the type of the front syllable when the front syllable is “Edan”. According to the dictionary search device, the dictionary is searched with a character string in which the long vowel symbol is uniquely converted into “i” or “e”.