JPH1069487A - Managing method for information with ambiguous name - Google Patents

Abstract

PROBLEM TO BE SOLVED: To input or retrieve information with an ambiguous name by providing a system, which registers information including a name and retrieves the information according to the name, with a dictionary wherein simplified names are made to correspond to formal names according to a rule based upon the features of the names. SOLUTION: An alias management table generation processing part 02 generates an alias management table 09 through an alias simplifying process 07 on the basis of a data base 08 having a formal management name. An information registering process part 02 inputs information and finds a formal name, and registers the input information in an information data base 10 together with the formal name. The information data base 10 is provided with a field for managing the formal names for a data base storing information and a field for managing ambiguous names that user have inputted. A name registration assist agent 04 searches for and outputs a formal name for an input having no formal name in registered information, and an information retrieving process part 05 retrieves information on the basis of the inputted name.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of managing information by an ambiguous name which is effective when a large number of people register and search by an ambiguous name in a database system on a computer system.

[0002]

2. Description of the Related Art When information is managed by a data processing system, an identifier for identifying the information is required. When using a name as an identifier, the person who enters the information and the person who searches for the information must know the name uniquely. When the number of names is large, or when it is difficult to predict the names to be used, it is difficult to correctly input names as identifiers, so that a dictionary or table of similar names is required.

There is a method of retrieving information having a synonymous notation by unifying the fluctuation of the notation (for example, JP-A-62-290968 and JP-A-62-290967).
JP-A-60-150176, JP-A-60-150176
No. 149630). Although it is possible to unify katakana and alphabet notation, it is not possible to unify specific names such as company names.

In order to search for information having similar names, there is a method in which an input name is divided into words and a similarity is searched from a dictionary (for example, Japanese Patent Application Laid-Open Nos. 3-41560 and 6-2311).
No. 88, etc.). If the name cannot be separated into words, or cannot be separated, a matching name,
I could only search for the names included.

[0005] Therefore, it cannot be solved by a dictionary or a table, and a dedicated inputting person for a name is required. For this reason, the person who acquired the information could not immediately input. Or, even if you could enter it, other people could not see it immediately.

[0006]

SUMMARY OF THE INVENTION The above conventional method has the following problems.
There is no consideration for the immediacy of inputting and sharing information by a large number of people, and there is a problem in processing names as identifiers.

An object of the present invention is to provide an identifier in two stages,
An object of the present invention is to provide a method for managing information based on an ambiguous name so that information can be input and searched from an ambiguous name.

[0008]

The above object is achieved by managing names in two stages and maintaining a dictionary of names (alias table). The two-stage management of names is performed as follows.

[0009] A name unification rule is determined, an alias is created based on the rule, and registered in the management name alias table.

[0010] A column for a name entered by a registrant and a name for information management is provided in the information, and a management name is automatically associated with the input name by referring to the alias table. The correspondence between the input name and the management name is divided into immediate processing and agent processing to shorten the information registration time.

For an input name that does not correspond even in the agent processing, a candidate is output to help registration in the alias table.

The information manager registers the input name without the management name in the alias table, or informs the input person of the error of the input name to correct it.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of a method for managing information by fuzzy names according to the present invention. The information management processor 01 includes an alias management table creation processor 02, an information registration processor 03, a name registration assistant agent 04, an information search processor 05, an information update processor 06, and a name simplification processor 07. You. The circled numbers on the lines from 02 to 06 indicate the order of processing.

The alias management table creation processing unit 02 shown in FIG. 1 creates an alias management table 09 by name simplification processing 07 based on a database having a formal management name (name database 08).

In the case of the name "Hitachi, Ltd."
Read "Hitachi, Ltd."
Pass to 7. The name simplification processing unit 07 returns the simplified name “Hitachi” to the alias management table creation processing unit 02. The alias management table creation processing unit 02 registers “Hitachi, Ltd.” and “Hitachi, Ltd.” in the alias management table 09 as a pair of an official name and an alias. Alias management table 09
2 is shown in FIG. 2, and the processing of the name simplification processing unit 07 is shown in FIG.

The table shown in FIG. 2 has a name simplification processing unit 07.
An abbreviation such as 104 is also registered in addition to the alias created by.

In the flowchart of the name simplification processing unit 07 in FIG. 5, aliases for the official names 103, 105 to 109 in FIG. 2 are created. This embodiment is a case of a rule for simplifying a customer name. The alias of 103 in FIG. 2 is created by the processing of 404 and 408 in FIG. Another name for 105 is 4
13 processing, 106 processing of 403, 107 processing of 407 and 412, 108 processing of 410, 109 processing of 415
Make by processing.

The information registration processing unit 03 shown in FIG. 1 inputs information, finds a formal name corresponding to the input name, and registers the input information in the information database 10 together with the formal name.

If the input information contains an input name of "Hitachi Computer", "Hitachi Computer" is passed to the name simplification processing unit 07. The name simplification processing unit 07 performs simplification processing and returns “Hitachi Computer”. The information registration processing unit 03 searches the alias management table 09 for an alias that matches “Hitachi Computer”, and finds the official name “Hitachi Computer” corresponding to the matched alias.
Take out. The official name “Hitachi Computer” and the input information are registered in the information database 10. Information registration processing unit 0
FIG. 3 shows an example of the information database registered in step 3.

The information database has a 201 field for managing a formal name for a database for storing information, and a 2nd field for managing an ambiguous name input by a user.
02 fields are provided.

FIG. 4 shows the operation of the formal name correspondence processing in the information registration processing section 03 of FIG. When adding or changing a record in the information database, the input ambiguous name is subjected to a name simplification process 400 (FIG. 5). The simplified name of the processing result and 10 in the alias management table
Compare the contents of 2. If they are the same, the official name of 101 corresponding to the same is registered in the 201 field of the information table. If the same thing is not found in 102, a thing containing the same contents is searched for from 102. If there is only one, it is registered in the field 201 as the official name. If nothing containing the same content is found, leave blank.

According to the information 203 in FIG. 3, “Hitachi, Ltd.” is simplified to “Hitachi, Ltd.” in the process of 404 in FIG. 5, and a match is found in the 102 column of FIG. 10
Since four lines were found, the corresponding official name "Hitachi, Ltd." is registered as the official name.

The name registration assistant agent 04 in FIG. 1 further searches for an official name having no official name in the registration information, and as a result (found official name, official name candidate, not found) (See FIGS. 6 and 7 for details). Information registration processing unit 03 and information update processing unit 06
Thus, the processing of the name registration assistant agent 04 is not included, with emphasis on the processing speed. Name registration assistant agent 0
4 is automatically started at a fixed time. Based on the execution result of the name registration assistant agent 04, it is determined whether or not the administrator registers an alias and whether to correct the input name.

FIGS. 6 and 7 show the operation of the name registration assistant agent 04 of FIG.

First, a record without a formal name (blank) is extracted from the information database. The ambiguous name 202 part of the record input by the user is simplified by name simplification processing 400, and the comparison and registration processing of 600 (FIG. 7) is performed. This processing is the same as the processing of the name registration processing 300, but the registration is performed when a new official name is added to the alias management table 101 in FIG. 2 or an abbreviation is added. This is the process. Next, the part enclosed in parentheses is deleted in the processing of 504,
This is simplified by name simplification processing 400, and processing 600 is performed.

Next, the contents of the input name of 202 extracted in the process of 507 are set so that a delimiter character “,. The name simplification process 400 is performed on this, and the process 600 is performed. Repeat until no more delimiters are found. Then 5
The contents of the input name 202 extracted in the process 07 are converted by the process of name simplification 400, and the last character is deleted.
00 is performed. This is repeated until the number of characters reaches an arbitrary number. If the corresponding formal name is not found after these processes, a report to that effect is made as in 514. 6
If a plurality of candidates are found in the process 605 of 00, the candidate name is notified to the administrator in the process of 606. If there is only one match and one candidate, the registered contents are reported to the administrator as indicated by 608. This is because the result of the comparison with the agent processed is not always correct.

For example, if the input name is "Hitachi (HITAC) Computer" or "Hitachi Computer: HIT
If “AC” is designated, it is converted to “Hitachi Computer” in 504 (or 507) and 505 (or 50).
8) Simplify to "Hitachi Computer" in 601-60
Register as the official name "Hitachi Computer" in 3. The result is notified at 608. When “Hitachi Biotechnology” is specified in the input name, 511 to 513 in FIG.
When searching the table of 102 while cutting the last character in the processing of, the character string is included in "Hitachi Violin" at the time of "Hitachi Bio", so "Hitachi Violin" is registered as the official name, The notification is made at 608, but the selection is incorrect and the alias management table 09 needs to be changed.

The information search processing unit 05 shown in FIG. 1 performs information search processing based on the input name. FIG. 8 is a diagram illustrating a search operation of the information search processing unit 05 in FIG. The input ambiguous search name is converted to a name simplification process 400 (FIG. 5).
Process. The simplified name of the processing result,
The contents 102 of the alias management table 09 are compared. If they are the same, a search is performed using the formal name of 101 and the input name. When "Hitachi Technology" is specified as the input name, "Hitachi Engineering Department" is selected as the formal name in the processing of steps 701 and 704, and the formal name of 201 and the user of 202 are selected using the formal name and the input name in 703. -Search for the name entered by the user.

As a result of the search, there is a hit at 210 in FIG.
If "Hitachi" is specified as the input name, the name including "Hitachi" is searched for 201 and 202 in the process of 705, and all 203 to 210 are hit.

The information update processing section 06 in FIG. 1 performs information update processing. If the input name is changed in the update process,
A name registration process is performed in the same manner as the information registration processing unit 03.

According to the present embodiment, when registering information, it is not necessary to be aware of the formal name, so that there is an effect that it is easy to input. Also, in the information search when various input names are registered, the name simplification process 07 and the name registration assistant agent 04 find the formal name, thereby increasing the information hit rate.

[0033]

According to the present invention, when managing information with an ambiguous name that is input by a large number of people, the information is made easier to input by associating the management name with the registered input name in the information, and the input name. By searching for the management name, the hit rate of information can be increased.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a system that manages information using ambiguous names.

FIG. 2 is a diagram illustrating an alias management table (dictionary) in which aliases, abbreviations, and the like by name simplification processing are registered in order to associate a management name (formal name) with an input name.

FIG. 3 shows a formal name,
FIG. 9 is a diagram showing a state in which the input name is added.

FIG. 4 is a diagram showing a process for registering a formal name with respect to an ambiguous input name during information registration.

FIG. 5 is a diagram showing the contents of a process for simplifying a name input by a user to make it easy to find a formal name.

FIG. 6 is a diagram showing the processing content of an agent for performing another process again and registering an input name for which a formal name was not found.

FIG. 7 is a diagram showing the contents of a sub-process used for registration in the process of FIG. 6;

FIG. 8 is a diagram showing a process of searching for and searching for a formal name from an ambiguous input name during information search.

[Explanation of symbols]

 01: Information management processing unit 02: Alias management table creation processing unit 03: Information registration processing unit 04: Name registration assistant agent 05: Information search processing unit 06: Information update processing unit 07: name simplification processing unit 08: name database 09: alias management table 10: information database 300: name registration processing 400: name simplification processing 500: name registration assistant processing

Claims (3)

[Claims] In a system for registering information including a name and retrieving information based on the name, a dictionary in which a simplified name is made to correspond to a formal management name in accordance with a rule capturing characteristics of the name. A method of managing information using ambiguous names, characterized by having 2. A system for registering information including a name and retrieving information based on the name, in the information registration process, simplification of the name of the input name and searching for a match from a dictionary in which the name is simplified. Information management by fuzzy names, which has a formal management name search process, further has a name registration assistant agent, and has a process of searching for an formal name after registering information for an input name for which no formal name was found Method. 3. A system for registering information including a name and retrieving information based on the name, wherein the information to be stored has an input name at the time of registration and a formal management name, and a search designation name is used as it is when retrieving information. And a process for retrieving an input name and a management name in the information based on the simplified name.