JPS6049931B2 - Information search method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an information retrieval method, and more particularly, to an information retrieval method for obtaining correct retrieval results from aural information or phonetic character information that includes homonymized voiced sounds in Japanese.

With the progress of the information society, a wide variety of search systems have appeared.

Particularly recently, with the rise of computer use and sumo wrestling, search systems are emerging that allow ordinary people to enter and search information. In this case, it does not mean that the general public operates directly. A typical example of this system is a computerized telephone directory assistance system.

That is, the information to be searched in this system is registered information from the telephone owner, and the search information is the offer information from the telephone number inquirer. The kana reading name is used as key information for these search targets and searches. In this case, there is a problem with the vagueness of the Japanese phoneticization rules. That is, when ideograms are converted into phonetic characters, a plurality of different expressions are commonly used for new and old kana usage, homophone notation, etc. One of these is the problem of 1 ji, ji/zu, and zu.

1 ji, ji Jrs, zu. The principles of the notation method are established in 1 modern Kanazukayo.

1.Modern Kanazukayo is a set of rules for writing modern words in kana based on modern speech sounds, and is mainly applied to colloquial sentences.

However, there are many cases in which words are written without following this rule, or in a vague manner.

For example, it is difficult to accurately write 1 ji, jo and 1 zu, zuo, such as Asaka as 1 Azumi Jr. Azumyo, Kazuo as 0 Kazuo 1 Kazuo, Miyaji as 1 Miyaji/Miyajo, etc. It is allowed to be written in
Therefore, in the telephone directory assistance system, the kana reading name (registered information) from the phone owner registered as searched information and the kana reading name (offer information) obtained by converting the offer from the phone number inquirer into phonetic characters. There are many cases where they do not match.

At this time, if these pieces of information are used as the searched key information and the search key information, the two will not match and a search solution will not be obtained. In order to eliminate this kind of inconvenience, it would be possible to establish a rule to match the searched information with the searched information, but since registered information such as person's name and business name is a matter related to the individual subscriber, it is not possible to force it to be unique. It is difficult to unify.

Generally, as a countermeasure for such a case, (1) prepare multiple search key information for the same subscriber corresponding to each kind of kana reading of the name; or (2) search key information. When searching for a name, a method is used to obtain a matching solution (search solution) by creating multiple search keys corresponding to each type of kana reading of the name.

However, in the case of (1) above, the searched information is accommodated. The file capacity increases, and in the case of (2) above, the search operation is troublesome or takes a long time.

The present invention eliminates the above-mentioned drawbacks, and makes it possible to complete the search in a short time without increasing the capacity of the file containing the searched information, and without increasing the number of search keys at the time of search. The purpose of this invention is to provide a search method for searched information having the following functions.

The purpose of the present invention is to provide an input device capable of inputting Japanese phonetic character strings, and to provide an input device capable of inputting Japanese phonetic character strings, as well as diacritic characters that are difficult to accurately represent in the phonetic character string input to the input device. A conversion means for converting the clarified characters into predetermined characters for the above characters to obtain key information, and an information file that uses the key information as the searched key information of the searched information. means for searching the information file for searched information having searched key information that matches the key information, using the key information obtained by converting the character string input for the search by the conversion means; and,
An output device that outputs the information searched from the above file, and a means for comparing the searched information with the character string input for the above search, and transferring the one with a good match to the above output device. In the search, the key information obtained by converting the character string input into the input device by the converting means is used to search the file, which has key information to be searched that matches the key information. This can be achieved by extracting the searched information, collating the extracted searched information with the character string input to the input device, and sequentially notifying the output device to output the strings with the best match. It will be done.

Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram of an example in which the present invention is implemented in a telephone directory assistance system.

In the figure, 1 is an input device for inputting information, and 2 is a character string input from the input device 1 that detects the corresponding characters (chi, ji, tsu, zu) included after the target character n in the character string inputted in advance. A conversion device 3 for converting into predetermined characters is a buffer memory 4 that stores character strings converted by the conversion device 2 and further converted into clear-sound characters by a control device 4 to be described later. It is a control device, and has a function of converting the character string converted by the conversion device 2 into clear-sound characters and setting it in the buffer memory 3, a function of creating searched information and setting it in the file storage device 5, and a function of the buffer memory 3. A plurality of pieces of searched information (search solutions) are extracted from the file storage device 5 using the character strings stored in the input device 1, a match with the character strings stored in the input device 1 is detected, and the search solution is determined based on the degree of matching. 5 is a file storage device that stores a large amount of information to be searched; and 6 is an output device that outputs weighted information extracted from the file storage device 5 by the control device 4. .

Now, I am creating the searched information file. This will be explained using an example.

In FIG. 1, first, the above n is set to 2 in the character position setting counter 7 of the conversion device 2.

Generally, the first character of declaration information rarely requires this type of processing, so set 2 here and enter the second character.
This is to perform processing starting from characters. However, when processing from the first character, set 1 as n, and
When processing starts from the 4th character, 3, 4, etc. are set. The information to be searched is input into the input device 1, and the input kana reading portion (in this case, Madda Kazuo corresponding to the name) is taken out and set in the buffer memory 9 of the conversion device 2.

The value 2 set in the character position setting counter 7 is set in the character number counter 8.

According to the value 2 set in the character position setting counter 7 from the beginning of the character string (Mazda Kazuo) set in the buffer memory 9, in this case, the second character is taken out, and the voiced sound conversion table 10 is searched to find the corresponding character. If it is, that is, ji, ji, tsu, zu, it is converted to shi, ji, su, zu, and stored in the buffer memory 9.
It is set at the position indicated by the numerical value of the contents of the character number counter 8, but if it does not correspond, nothing is done.

In this case, since the numerical value of the content of the character position setting counter 7 is 2, the second character 0tsuo is extracted and the voiced sound conversion table 1
When 0 is searched, since it corresponds to the above character, it is converted to a ``s'' and this is set in the second position determined by the numerical value of the contents of the character number counter 8, in this case 2. Here, 1 is added to the contents of the character number counter 8.

The above process is repeated until the above check is completed for all characters of the character string set in the buffer memory 9 in this way. When this process is completed, the character string 0 Mazda Kazuo, which was initially set in the buffer memory 9, is converted to 1 Masuda Kazuo.

All of the above controls are performed by the conversion control device except for setting 2 to character position setting counter 7 and inputting
This is done under the control of

Furthermore, the character string 1 set in buffer memory 9
The Masuda Kazuo is converted into clear sound characters by the control device 4 and set in another buffer memory 3.

Next, the control device 4 converts the character string set in the buffer memory 3 into the information stored in the input device 1 as searched key information of the searched information stored in the input device 1. It is additionally stored in the file storage device 5.

That is, the above information is stored in the file storage device 5 as shown in FIG.

In the figure, K is a searched key information section, and D is a searched information section. When one piece of declaration information is stored in a file as described above, the contents of the input device 1, buffer memories 3 and 9, and character number counter 8 are cleared and initialized to prepare for the next input operation. Next, when the following declaration information is given as searched information, it is stored in a file in the same manner as above.

Then, each piece of searched information is rearranged in the order of the kana reading information of the searched key information, and a searched information file is created.

In this case, the contents of the file will be as shown in FIG.

Next, a process of searching for necessary information from the information file created as described above will be explained. An explanation will be given using one Mazda Kazuo as an example of the declaration information (interview information) from the number inquirer. In this case, the character position conversion counter 7 is set to the same number 2 as when the file was created.

When the character string "Mazda Kazuo" is created based on the listening information 1 "Mazda Kazuo" and inputted into the input device 1, it is first transferred and set in the buffer memory 9 of the conversion device 2,
Furthermore, the same process as described above is performed, and the character string 1 is set in the buffer memory 3. Here,
The control device 4 reads the character string 1 set in the buffer memory 3.
Using the master search key as a search key (primary key), from among the searched information stored in the file storage device 5, three items shown in Table 1 are found in which the above search key and the searched key of the searched information match by a known method. Extract.

Next, the control device 4 uses the declaration information stored in the input device 1 as a secondary search key, and compares it with the kana reading part of the extracted data, and as a result, the above secondary search key information. The output device 6 gives priority to those whose pronunciations have a large number of consecutive characters that match the beginning over others.
to notify.

That is, the output device 6 is notified in the order shown in Table 2 below. The sort search is completed when the output device 6 outputs the notified solution.

In addition, in the above search, if you input a character string 0 Mazda Kazuo, which is different from the above, based on the interview information 1 Mazda Kazuo, the input character string will be Conversion device 2
The character string 1 is converted into the same character string 1 as above and set in the buffer memory 3, a search is performed according to this, and the searched information shown in Table 1 is extracted in the same way as above.

Then, the priority order is determined by comparing the information stored in the input device 1 (secondary search key) with the kana reading part of the searched information. The ranking will be as shown in Table 1, and the correct answer will be notified to the output device 6 second. In the above, from the character string input to the input device 1, only one searched information key and the same search information key as the one and only searched key must be created.

For example, the above-mentioned 1 Miyajo may be written as 1 Miyajo, or 1 Miyajo outside of 1 Miyajo.

If it is input as 1 miyajo or 1 miyajo, it is converted to 1 miyajo and set in the buffer memory 3, as is easily understood from the above.

Also, when input as 1 Miyacho, voiced sound conversion table 1
1 cho is converted to 1 cho by 0, and 1 cho is the same as above.
It is set in the buffer memory 3 as key information. For example, in the case of 1 kotsukyo, it may be written in three ways: 1 kotsukyo, 1 kozukyo, 1 kozukyo, but if 1 kotsukyo is input, fu becomes 1 kozukyo according to the voiced sound conversion table 10. .

If ``kouzukyo'' is inputted, it will be converted to ``kouzukyo'' according to the voiced sound conversion table 10, and then it will be purified in the control device 4 to become 1 kouzukyo. Although it is not converted, it is cleared in the control device 4 and converted to 0 kosuki. In both cases, the same key information is set in the buffer memory 3 as 1 kosuki. It can be constructed as a conversion table that converts ``noshi'' to ``su'' and tsu, zu to ``su''.

Also, conversely, 1 sho 1 jo 1 s J 1 zuo can be changed to 1 ji Jr./jo 1 zuo, or 1 shi, jo is 1 jo.
It is possible to create a conversion table that converts cho to Jsu and zuo to 1 tsho. As explained above, according to the present invention,
5 dakuon characters that are difficult to write correctly when aural information or ideographic information is converted into phonetic characters (kana characters),
In a Japanese information processing system that handles phonetic character strings including も1zu and ㅔ, it is possible to easily search without increasing the file size or the number of search keys at the time of search.

In addition, the expansion of search solutions created by converting dakuon characters, which are difficult to display correctly, and their clarified characters into predetermined characters, gives importance to the information provided by the inquirer. This can be easily resolved by prioritizing the issues.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2 is a diagram showing a memory file containing one piece of searched information, and Fig. 3 is a block diagram of an embodiment of the present invention.
The figure shows a memory file containing a plurality of pieces of searched information having the same searched key. 1... Input device, 2... Conversion device, 3, 9... Buffer memory, 4... Control device, 5... Memory file, 6... Output device, 7... Character position setting Counter, 8... character number counter, 10... voiced sound conversion table,
11...Conversion control device, K...Searched key information section,
D...Searched information section.

Claims (1)

[Claims] 1. An input device capable of inputting Japanese phonetic character strings, and an input device that can input voiced characters that are difficult to accurately represent in the phonetic character strings input to the input device, and their clarified characters. a conversion means for converting the above-mentioned characters into predetermined characters to obtain key information; a storage device that configures an information file with the key information as the searched key information of the searched information; and an input search means for searching the information file for searched information having searched key information matching the key information from the information file using the key information obtained by converting the character string for and a means for comparing the retrieved information with the above-mentioned search character string and transmitting the one with a good match to the above-mentioned output device. , Extract searched information having searched key information that matches the key information from the file using the key information obtained by converting the character string input into the input device by the conversion means, and further extract the searched information from the file. An information retrieval method characterized in that the retrieved information that has been searched is compared with a character string input to the input device, and the information that is a good match is sequentially notified to an output device for output.