JPH07200256A - Method for sorting word irrespective of language - Google Patents

Abstract

PURPOSE:To digitize languages for smoothly sorting the words of the languages without depending on the vague memory of human as much as possible. CONSTITUTION:The word cannot be sorted as it is. For sorting the word, KANJI(Chinese character), KANA(Japanese syllabary), KATAKANA(square form of KANA), and foreign languages(alphabets, Arabic script and the like) are digitized in accordance with a situation so as to compare them. (Sorting can be realized by digitizing the language to be sorted in accordance with a definition file and by making it possible to be measured.) Only when numbers corresponding to the words do not exist, or when pronunciation different from a regular case is executed, learning is executed by a manual operation. Thus, sorting corresponding to a designated condition can be executed in a form where errors and labor are required minimum irrespective of the language.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to sorting of notations (kanji, katakana, hiragana, alphabet, etc.) input as general data.

[0002]

2. Description of the Related Art Generally, there are various sources of data that need to be sorted, such as a business card recognition device that has recently become popular, an image scanner, an online database such as a personal computer communication, etc., and other means for manually inputting information. Although there is information peculiar to characters in the information entered by, there is no reading name or pronunciation information. Therefore, sorting in the order of AIUEO was a fairly difficult task.

[0003]

In such a case, even if one sort in Japanese aiueo is taken, it takes a lot of labor and efficiency is never good. When sorting Japanese characters in aiueo order, if kanji or alphabets are included, convert them into kana once for comparison, and if they are applied to a computer, perform some encoding for comparison. It will be. However, Japanese kanji and foreign languages sometimes have multiple kana names, which is not easy but difficult to say.
In addition, if you have to sort with complicated conditions, or sort things that meet conditions such as a set based on a person's name, you will have to rely on people's judgments one by one, and the computer will be difficult to use.

Further, the sorting is not limited to the Iueo order based on Japanese, but may be the Irohanihohet order or the like, and the standard changes variously.

When inputting data, it is also possible to input information (codes) for performing sorting in parallel, but since the resources of the storage medium are deprived by that much,
I don't think it's a good way.

[0006]

DISCLOSURE OF THE INVENTION The present invention is to enable sorting of languages that cannot be easily sorted without using human hands as much as possible. There are various sorts of sorting methods for encoded comparable groups, such as bubble sort and insertion sort. However, there are merits and demerits depending on the number of materials, so the sorting method should be changed according to the situation. Will change. The problem is how to put what can't be sorted into a sortable state. This is achieved by digitizing the languages you want to sort and making them measurable without relying on human resources as much as possible.

When sorting words, kanji, hiragana, katakana, and foreign languages (alphabets, Arabic letters, etc.) are digitized according to the situation so that they can be compared. Learn only if the word does not have a corresponding number or if it sounds different than usual.

[0008]

[Example] When attempting to sort a group of words in Table 1 (personal name as an example), if the descriptions in Hiragana are not done in parallel, the human labor is necessary for sorting. Become.

[Table 1]

However, one of the following three characteristics applies to how to read a word including Japanese Kanji.

"Osaka" is "Ohan" and "Daisaka"
Is not read. "Osaka" is read only by "Osaka". Cases in which such definite reading is decided (definite reading) Cases in which reading is often carried out in an alternative way such as "Kiko" or "Noriko" or "Masako" (selective reading) ) A difficult case (unpredictable or esoteric reading) that causes an unpredictable reading that makes "conquest" read as "Masano"

[0011] can mechanically convert "Kanji to Hiragana". Only in cases such as or, the conversion is performed manually as unconvertible. , Are defined as dictionaries so that kanji and hiragana are paired as in “Table 2” and “Table 3”.

[Table 2]

[Table 3]

Furthermore, the conversion of "Kanji to sign" for the family name is a sort condition.

Perform according to the definition in [Table 4]. (The code uses a decimal number for the sake of clarity.) This conversion method may use two methods (|, ‖) depending on whether is frequently used.

[Table 4]

| Different words in dictionary are frequently used. Dictionary is pre-encoded and directly encoded without compromising the elements of the set to be converted. ‖Dictionary words are not frequently used in duplicate. Use a dictionary to temporarily make the elements of the set to be sorted into kana. Encoding is performed on what is kana.

Whichever method is used, the result is as shown in Table 5.

[Table 5]

Further, in order to make a comparison including a space, the rearrangement according to the longest code is performed, and the result is shown in Table 6.

[Table 6]

When the name is included, it becomes as shown in "Table 7".

[Table 7]

Further, in order to make a comparison including spaces, the rear end is arranged according to the longest code, as shown in "Table 8".

[Table 8]

When sorting is performed based on the created codes, the result is shown in "Table 9".

[Table 9]

It is possible to sort words including Kanji at least in the above-mentioned form. When it becomes necessary to register a word in kana, it will be added to the dictionary as needed.

When adjustment is required, such as when it is determined to be erroneous reading, the user changes the definition of the dictionary as necessary, or the system that provides this encoding responds to the content of the adjustment. Deal with it by changing the dictionary. In this sort, "Kyo" and "Gyo",
The definition is made so that "kiyo" and "giyo" are completely separated, but it is possible to make any change by changing the file ("Table 4" in the actual example) necessary for digitizing. (A configuration in which one character is paired with a plurality of characters is also possible.)

[Table 4]

[0021]

[Effect] Sorting can be performed according to specified conditions in a form that requires the least error and effort regardless of the language.

Claims (2)

[Claims] 1. A method of converting a set of Japanese and foreign languages including kanji, katakana, etc. into a code necessary for sorting according to a specified condition, and sorting from the order of the code. 2. When converting to a code necessary for sorting the preceding claims, when there is no corresponding code, or
When the conversion is judged to be a wrong code, the error is corrected and the definition is changed if necessary.