JPH0341560A - Information retrieving system - Google Patents

Abstract

PURPOSE:To efficiently retrieve information having similar names by referring to a dictionary to obtain the representative word of words constituting a name specified by an operator. CONSTITUTION:A representative word conversion part 8 divides a name inputted from a name registering part 7 into several words while referring to words registered in a word storage part 3 and obtains the representative word ID of a representative word corresponding to respective words. Namely, the name is registered 6 by referring to the dictionary, dividing the name to be registered into words, obtaining the representative word corresponding to the divided words and forming and storing a list indicating the correspondence between the name and the representative word. At the time of executing retrieval, the name including the same representative word as the representative word of words constituting the name specified by the operator is retrieved from the list. Consequently, a similar name corresponding to the name specified by the operator can be efficiently obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an information retrieval method, and particularly to an information retrieval method suitable for retrieving a name similar to a name specified by an operator.

[Conventional technology]

Conventional search systems have a dictionary in which words are registered, and searches for similar names are performed by searching for names that include the same words as the words that constitute the name specified by the operator.

[Problem to be solved by the invention]

The above conventional technology is based on a dictionary in which words are registered, and when searching for a name similar to the name specified by the operator, only those that include the words that make up the name are obtained, and although they are used for the same purpose. There has been a problem in that names that include words with different spellings (for example, use and utilization, telephone number and phone number, etc.) cannot be searched as similar names.

An object of the present invention is to provide an information search method that allows searching for similar names that include similar words written differently.

[Means to solve the problem]

In order to achieve the above object, the present invention registers a representative word representing several words used with the same intention in addition to words constituting one name in a dictionary for registering words, and when registering a name, A list is generated and maintained that shows the correspondence between a name and the representative words of the words that make up the name, and when a search is executed, the dictionary is referred to and the representative words of the words that make up the name specified by the operator are retrieved. In addition, a name including the same representative word as the representative word is obtained from the list.

[For production]

To register a name, refer to a dictionary, divide the name to be registered into words, obtain representative words corresponding to the divided words, and generate and maintain a list showing the correspondence between names and representative words. put. When executing the search, a similar name corresponding to the name specified by the operator is obtained by searching the list for a name that includes the same representative word as the representative word of the word constituting the name specified by the operator.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of one embodiment of the present invention.

In the figure, 1 is a terminal device operated by an operator that registers words that make up a name and representative words, 2 is a word registration unit that registers words that make up a name, and 3 is a word that saves the registered words. A storage unit, 4 is a representative word registration unit for registering representative words, 5 is a representative word storage unit for saving registered representative words, 6 is a terminal device operated by an operator who registers names, and 7 is a name storage unit. 8 is a name registration unit that performs registration; 8 is a representative word conversion unit that converts words constituting a name into representative words; 9 is a name storage unit that stores registered names and the correspondence between names and representative words; 10 is a name storage unit that stores similar names. A terminal device 11 is operated by an operator who performs a search, and a search processing unit 11 performs a search for similar names. Note that although the terminal devices 1, 6, and 10 are shown separately for convenience, they may be the same.

The operations shown in FIG. 1 are broadly divided into registration and search, and registration is divided into registration of words and representative words, and registration of names.

When registering words and representative words, the operator inputs the words and representative words together with their IDs from the terminal device 1. The word registration section 2 stores the words input from the terminal device 1 in the word storage section 3.0 The representative word registration section 4 stores the representative words input from the terminal device 1 in the representative word storage section 5.

When registering a name, the operator inputs the name and its name ID from the terminal device 6. The name registration unit 7 is the terminal device 6
The input name is sent to the representative word conversion section 8. The representative word conversion unit 8 divides the name input from the name registration unit 7 into several words by referring to the words registered in the word storage unit 3, and refers to the representative words registered in the representative word storage unit 5. Then, obtain the representative word ID of the representative word corresponding to each word.

This representative word ID is returned to the name registration unit 7 together with the representative word. The name registration unit 7 stores the correspondence between the name input from the terminal device 6 and the representative word such as the representative word ID returned from the representative word conversion unit 8 in the name storage unit 9.

When executing a search, the operator inputs a search target name from the terminal device 10. The search processing unit 11 sends the name input from the terminal device 10 to the representative word conversion unit 8.

The representative word conversion unit 8 divides the name input from the search processing unit 1 into several words by referring to the words registered in the word storage unit 3, and converts the name registered in the representative word storage unit 5 into several words. Refer to the representative word ID of the representative word corresponding to each word.
This representative word ID is then returned to the search processing unit 1. The search processing section 11 obtains the name corresponding to the representative word ■D returned from the representative word conversion section 8 from the name stored in the name storage section 9, and sends the search result to the terminal device 10.

FIG. 2 shows the format of the contents in the word storage section 3, representative word storage section 5, and name storage section 9, and their pairwise relationships. The word storage unit 3 holds a word ID, a word, and a representative word ID pair for each word. Representative word storage section 5
holds a representative word ID, representative word, type, and representative word ID string for each representative word. An abbreviation or NULL is registered as the type of representative word. The representative word whose abbreviation is registered indicates that it is an abbreviation of a string of several representative words, and in this case, the representative word ID column of the original representative word is registered in the representative word ID column. . In addition, the type is NU
The representative word ID column of the representative word in which LL is registered is NULL.
It is. The name storage section 9 consists of a name correspondence section 9-1 and a name search list 9-2.

The name correspondence section 9-1 holds a name ID, a name, and a representative name pair for each name, and the name search list 9-2 holds a name ID and a set of a representative name.
and a set of representative word IDs.

The words and representative words to be registered from the terminal device 1 are
When obtained together with D, the word registration unit 2 obtains the word ID, the word,
The representative word ID is registered as one set in the word storage unit 3, and the representative word registration unit 4 stores the representative word ID, representative word, abbreviation, etc. in the case of its type, and ID string of the original representative word as one set. Register it in the word storage section 5.

Upon obtaining the name to be registered along with its ID from the terminal device 6, the name registration section 7 sends the name to the representative word conversion section 8.

The representative word conversion section 8 first refers to the words in the word storage section 8, divides the name into several words, and obtains a representative word ID corresponding to each of the divided words. Next, obtain the representative word corresponding to this representative word ID from the representative word storage unit 5,
Along with the representative word, if the type of the representative word is NULL, the representative word ID is returned to the name registration unit 7, and if the type is an abbreviation, the contents of the representative word ID string are returned to the name registration unit 7. The name registration unit 7 registers a set of representative names generated by combining the name ID and name input from the terminal device 6 and the representative words from the representative word conversion unit 8 in the name correspondence unit 9-1, moreover,
A correspondence list of name IDs and representative word IDs is registered in the name search list 9-2.

FIG. 3 shows a specific example of the contents of the word storage section 3, representative word storage section 5, and name storage section 9. Hereinafter, a specific example of the search operation will be explained based on this.

Now, assume that the name "Customer Electric Power" is input from the terminal device 10. The search processing section 11 sends this "customer@den" to the representative word conversion section 8. The processing in the representative word converting section 8 is basically the same as when the name to be registered is given from the name registration section 7. First, with reference to the word storage unit 3, "Customer Denden" is divided into "Customer" and "Electric Denden", and the representative word IDs corresponding to each of the divided words are "1ko" and "15".
” to obtain the representative word ID of this “11” and “15”
With reference to the representative word storage unit 5 as a key, "customer", "
Obtain a representative word for "Denden", use the corresponding representative word IDrllJ for "Customer", and use the representative word ID corresponding to the original representative words "Telephone" and "Number" since the type of "Denden" is an abbreviation. r12" and "13" are returned to the search processing unit 11.

In the search processing unit 11, first, the representative words ID rllJ, r12J returned from the representative word conversion unit 8 are used.

Refer to the name search list 9-2 of the name storage section 9 using "13" as a key, and select r21 as the corresponding name ID.
Obtain J, r22J, and r23J. Next, this name I
The name correspondence section 9-1 is referred to using Dr21J, r22J, and r23J as keys, respectively.

"Subscriber name", "Subscriber telephone number", and "User telephone number" are obtained as names having the name ID and sent to the terminal device 10.

As described above, "Subscriber name", "Subscriber telephone number", and "User Denden" are obtained as similar names to "Customer Denden".

〔Effect of the invention〕

As explained above, according to the information search method of the present invention,
It is possible to search for names that are used with the same intention but have different spellings, and it is possible to efficiently search for information with similar names.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the information retrieval method of the present invention, FIG. 2 is a diagram showing the content format and correspondence of the word storage section 3, representative word storage section 5, and name storage section 9, and FIG. 3 1 is a diagram showing a specific example of the contents of the Japanese language storage section 31, the representative word storage section 5, and the name storage section 9. 1... Terminal device, 2... Word registration section, 3... Word storage section, 4... Representative word registration section, 5... Representative word storage section, 6... Terminal device, 7.・Name Registration Department,
8... Representative word conversion section, 9... Name storage section, 10
...Terminal device, 1. ...Search processing unit.

Claims (1)

[Claims] (1) An information search method that searches registered names for names similar to the name specified by the operator, including words that make up the name and a representative word that represents several words used with the same intention. When registering a name, the above dictionary is referenced to generate and maintain a list showing the correspondence between the name and the representative words of the words that make up the name, and when a search is executed, the above An information retrieval method characterized in that a representative word of a word constituting a name specified by an operator is obtained by referring to a dictionary, and a name that includes the same representative word as the representative word is obtained from the list.