JPH1049550A - Data base retrieval device and program storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a specifying method for retrieval conditions using ambiguous expression by calculating an ambiguous item value that the ambiguous expression indicates when the ambiguous expression is inputted. SOLUTION: A generating means 12 generates a statistical information data base 11 once a data base 2 is set up. On receiving the statistical information data base 11 (subordinate statistical information data base 13), a calculating means 16 calculates the ambiguous item value that the ambiguous expression indicates from statistical information that the ambiguous item value of the retrieval condition managed by the statistical information data bases 11 and 13 and a set value allocated to the ambiguous expression managed by a managing means 15. After the calculation processing of the calculating means 16, a determining means 17 determines the retrieval condition by using the ambiguous item value calculated by the calculating means 16 as it is, and displaying it on a display screen for corrections through interactive processing when the retrieval condition is instructed by specifying an essential item value and the ambiguous expression, and a retrieval means 10 receives the retrieval condition to perform retrieval from the data base 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database search device for searching a database and a program storage medium for storing a program for realizing the database, and more particularly to a database search device capable of specifying ambiguous search conditions. And a program storage medium for storing a program for realizing the same.

[0002]

2. Description of the Related Art When a search condition is given, a database search apparatus selects, from data managed by a database,
A process for searching for data that satisfies the search condition is performed.

A conventional database search apparatus employs a configuration in which a search condition that is uniquely determined by a numerical value or the like can be specified, and a configuration in which search condition data specified in that format is searched. .

[0004] For example, when searching for a recruitment information database, if wages are used as search conditions, for example, wages of 500,000 yen or more are set as search conditions by interacting with the user, and the search conditions are met. It was processing to search for job information.

[0005]

However, if only a uniquely defined search condition is permitted as in the prior art, the user may not be able to easily search for the data he or she wants depending on the search condition. There was a problem.

For example, in the case of retrieving a recruiting information database, for example, a user does not know whether his or her input wage is high or low, and repeats the search process many times while changing the wage input. There was a problem that it was not easy to search for the job information you wanted.

The present invention has been made in view of the above circumstances, and provides a new database search apparatus that can solve the problems of the prior art by enabling specification of ambiguous search conditions. And a program storage medium for storing a program for realizing the above.

[0008]

FIG. 1 shows the principle configuration of the present invention. In the figure, reference numeral 1 denotes a database search device provided with the present invention, and 2 denotes a database to be searched by the database search device 1 of the present invention.

The database search device 1 of the present invention includes a search means 10, a statistical information database 11, and a generation means 1.
2, sub-statistical information database 13, and creating means 14
, Management means 15, calculation means 16, and determination means 17
And learning means 18.

The search means 10 searches data managed by the database 2 for data satisfying a search condition. The statistical information database 11 manages statistical information of each data group of the database 2 divided by the essential item values of the search condition. The generation unit 12 generates the statistical information database 11 from the data of the database 2 when setting up the database 2.

The sub-statistical information database 13 manages the statistical information of each data group of the database 2 divided by the values of the required items smaller than the required items of the statistical information database 11. The creating unit 14 creates the secondary statistical information database 13 from the data of the statistical information database 11.

[0012] The management means 15 manages prescribed setting values assigned to the ambiguous expressions specified for the ambiguous items of the search condition. The calculating unit 16 calculates an ambiguous item value indicated by an ambiguous expression specified for the ambiguous item of the search condition using the setting value managed by the managing unit 15.

The deciding means 17 calculates the calculating means 16 when the retrieval condition is designated by specifying the required item value and the ambiguous expression.
The search condition is determined by using the ambiguous item value calculated as described above as it is, or by displaying it on a display screen and modifying it according to the interactive processing. The learning unit 18 learns the set value managed by the management unit 15 using the correction value to be corrected by the determination unit 17 and the statistical information used to calculate the ambiguous item value of the correction source.

Here, the database search device 1 of the present invention
Specifically, the database search function of the present invention is realized by a program, and this program is provided from a medium, stored in the database search device 1, and operated on a memory, thereby realizing the present invention. Become.

In the database search device 1 of the present invention configured as described above, the generation means 12 generates the statistical information database 11 when the database 2 is set up. Then, when the required item specified by the search condition does not match the one used for generating the statistical information database 11 and becomes a part thereof, the creating unit 14 The statistical information database 13 is created.

The statistical information database 11 / sub-statistical information database 13 generated in this manner manages the statistical information of each data group of the database 2 divided by the essential item values of the search condition, thereby performing the search. It manages the statistical information of the ambiguous item value of the condition.

Receiving the statistical information database 11 (sub-statistical information database 13), the calculating means 16 specifies, for example, an essential item value and an ambiguous expression and specifies a search condition to the determining means 17. The statistical information having the ambiguous item value of the search condition managed by the statistical information database 11 (sub-statistical information database 13)
Then, an ambiguous item value indicated by the ambiguous expression is calculated from the set value assigned to the ambiguous expression managed by the ambiguous expression.

In response to the calculation processing by the calculation means 16, the determination means 17 directly specifies the ambiguous item value calculated by the calculation means 16 when a search condition is designated by specifying an essential item value and an ambiguous expression. By using it or displaying it on a display screen and modifying it according to the interactive processing, the search condition is determined.
0 searches the database 2.

When the determining means 17 determines the correction value, the learning means 18 determines the correction value and the statistical information (statistics of the ambiguous item value) used for calculating the ambiguous item value of the correction source. Information), a change value of the setting value managed by the management unit 15 is derived, and the derived value is recorded. The average value of the recorded data is calculated to learn the setting value managed by the management unit 15. I do.

As described above, according to the database search device 1 of the present invention, a method of specifying search conditions using an ambiguous expression such as “high” or “somewhat high” is possible, so that the user can You can easily search for the data you want from the database.

Then, the ambiguous expression is learned so as to match the user's sense, so that the user can accurately retrieve desired data from the database.

[0022]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail according to embodiments. FIG. 2 shows an embodiment of the database search device 1 of the present invention.

The database search device 1 of this embodiment is
The job information data 3 recorded on a floppy disk or the like is used as an input to interact with a user operating the terminal device 4 to search for job information desired by the user stored in the job information data 3, Search program 2
0, a setup program 21, a learning program 22, a job information database 23, a job statistics information database 24, and a sub job statistics information database 25.
, An ambiguity management table 26, and a logging file 27.

Here, the search program 20, the setup program 21, and the learning program 22 are provided from a medium. FIG. 3 shows an embodiment of a processing flow executed by the setup program 21, FIGS.
FIG. 6 shows an embodiment of a processing flow executed by the search program 20, and FIG. 6 shows an embodiment of a processing flow executed by the learning program 22. Next, the present invention will be described in detail according to these processing flows.

When the setup program 21 receives a request from the user to set up the recruitment information database 23, first, as shown in the processing flow of FIG. 3, the recruitment information data 3 is read and the recruitment information database 23 is set up. Then, this job information database 2
Then, a job recruitment statistical information database 24 is generated from the job information 3.

FIG. 7 shows the data structure of the job information database 23, and FIG. 8 shows an example of the job information database 24. The recruiting information database 23 manages recruiting information issued by a recruiting source.
As shown in (1), recruitment information including age range, wage range, place of employment, job content, and the like, with gender and occupation as key items, are managed.

On the other hand, the job recruitment statistics information database 24
It manages the statistical information of each data group of the recruitment information database 23 divided by the essential item values of the search condition, and is divided by the essential item values of gender / occupation / age range as shown in FIG. Recruitment Information Database 2
3 manages the number of data items in each data group, the average wage of job information belonging to the data group, and the standard deviation of wages of job information belonging to the data group. When it is not necessary to generate the sub recruitment statistical information database 25, the management of the number of data items can be omitted.

In this way, the setup program 21 sets up the recruitment information database 23 as shown in FIG. 7 upon receiving a setup request from the user from the recruitment information database 23, and outputs the requisition information database 23 from FIG. The job requisition information database 24 as shown in FIG.

On the other hand, the search program 20, when the user specifies a required item of the search condition prior to the input of the search condition, first, as shown in the processing flow of FIG. When it is determined whether the item is looser than the required item used for generating the job recruitment statistical information database 24, and when it is determined that the item is the same as the required item used for generating the recruiting statistical information database 24, When the process is terminated without performing any processing and it is determined that the required item is looser than the required item used for generating the recruiting statistical information database 24, that is, the specified required item is used for generating the recruiting statistical information database 24. When it is determined that the requisition is a part of the required item, the sub recruitment statistics information database 25 is generated from the requisition information database 23.

This sub job recruitment statistics information database 25
Manages the statistical information of each data group of the job information database 23 divided by the value of the designated required item. Like the job information statistical database 24 shown in FIG. It manages the number of data belonging to the group, the average wage of job information belonging to the data group, and the standard deviation of wage of job information belonging to the data group. The management of the number of data items can be omitted.

The sub job recruitment statistical information database 25 has some of the essential items of the statistical information database 23 as essential items. For example, in the job recruitment statistics information database 24 shown in FIG.
Although the age range is a required item, the sub-recruitment statistics information database 25 makes a part of the required item mandatory, for example, the gender / age range in the sub job recruitment information database.

According to this structure, the sub recruitment statistical information database 25 can be easily created from the recruitment statistical information database 24 without returning to the recruitment information data 3.

For example, in the case of the job recruitment statistics information database 24 shown in FIG.
The job information for "29 years old" has "98" data items, the average wage is "310,000 yen", the standard deviation is "58,000 yen", and "male, architectural design, 25-29 years old" In the recruitment information, the number of data is "28", the average wage is "350,000 yen", and the standard deviation is "620,000 yen". No, the average wage for job offers for "male, 25-29"

[0034]

(Equation 1)

According to the above, "320,000 yen" is obtained.
The standard deviation is

[0036]

(Equation 2)

Thus, "60,000 yen" can be easily obtained from the job recruitment statistical information database 24 using the equations (1) and (2). Using the nature of the sub recruitment statistics information database 25, the search program 20 generates the recruitment statistics information database 24 when the user specifies a required item of the search condition prior to input of the search condition. When it is determined that the required item is looser than the required item, the sub job-recruiting statistical information database 25 is generated from the job-recruiting information database 23 in accordance with a prescribed calculation process.

When the sub job recruitment statistics information database 25 is generated, the search program 20 performs the retrieving process using the sub job recruitment statistics information database 25 instead of the job recruitment statistics information database 24. For the sake of explanation, the sub-job statistics database 2
Assume that 5 is not generated. That is, it is assumed that the search program 20 performs a search process using the job recruitment statistical information database 24.

After the job statistics information database 24 has been generated in this way, when the user issues a search instruction for job information, the search program 20 first executes the steps shown in the processing flow of FIG. In step 1, a search condition input screen as shown in FIG. 9 is displayed on the display screen of the terminal device 4 to input search conditions from the user.

This search condition input screen employs an input method using vague expressions such as “high”, “somewhat high”, “normal”, “somewhat low”, and “low”. In step 1, for required items such as gender, age, and occupation, enter unique input values represented by numerical values, and for vague items such as wages, enter vague input values represented by vague expressions. Will do.

In step 1, the required item value and the ambiguous expression serving as search conditions are input. Then, in step 2, the average wage and the standard deviation indicated by the input required item value are retrieved from the recruitment statistical information database 24. get. For example,
When the job information database 24 manages data as shown in FIG. 8, "man, architect, 37 years old"
Is input from the job recruitment statistics information database 24, the average wage "38" indicated by the required item value is entered.
The data of "10,000 yen" and the standard deviation of "73,000 yen" are acquired.

Subsequently, in step 3, by referring to the fuzzy management table 26, the coefficient value indicated by the input fuzzy expression is obtained. The fuzzy management table 26 manages coefficient values assigned to fuzzy expressions. For example, as shown in FIG. 10A, the fuzzy expression that wages are “high” has a coefficient value “0.5”. ”, The coefficient value“ 0.3 ”for the ambiguous expression that the wage is“ slightly high ”, the coefficient value“ 0 ”for the ambiguous expression that the wage is“ normal ”, and the coefficient value for the ambiguous expression that the wage is“ slightly low ” The coefficient value “−0.5” is managed for the ambiguous expression “−0.3” and the wage “low”.

This coefficient value is specifically shown in FIG.
As shown in the figure, the coefficient is a multiplication coefficient for the standard deviation σ, and the coefficient value “0.5” indicates a wage higher by “0.5σ” than the “average wage”, and the coefficient value is “0.3”. Indicates a wage that is “0.3σ” higher than “average wage” and has a coefficient value of “0”.
Indicates the wage of “average wage”, and the coefficient value “−0.3”
The coefficient indicates a wage that is lower than “average wage” by “0.3σ”, and the coefficient value “−0.5” indicates a wage lower by “0.5σ” than “average wage”.

From now on, when the coefficient value is obtained from the fuzzy management table 26 in step 3,
Then, using the coefficient value and the average wage / standard deviation obtained in step 2, according to a calculation formula of wage = (average wage) + (coefficient value) × (standard deviation), the wage indicated by the input ambiguous expression is calculated. The calculated value is displayed on the display screen of the terminal device 4.

Subsequently, in step 5, it is determined whether or not the user enters a correction value for the display of the wage, and when it is determined that the correction value is not input, that is, If it matches the user ’s wage,
Proceeding to step 7, determine that the search condition specified by the mandatory item value input in step 1 and the wage calculated in step 4 is the search condition requested by the user, and the data satisfying the search condition Is retrieved from the recruitment information database 23, and the retrieval result is displayed on the display screen of the terminal device 4.

On the other hand, when it is determined in step 5 that the user inputs a correction value for the display of the wage indicated by the ambiguous expression, that is, when it does not match the wage which the user thinks, step 6 , The wage indicated by the ambiguous expression is changed according to the correction value input by the user, and the corrected value, the ambiguous expression input in step 1 and the average wage / standard deviation obtained in step 2 are recorded in the logging file 27. To be stored.

Then, the process proceeds to a step 7, wherein it is determined that the search condition defined by the corrected wage and the required item value inputted in the step 1 is a search condition requested by the user, and the search condition is determined. Is retrieved from the requisition information database 23 and the search result is displayed on the display screen of the terminal device 4.

In this way, the search program 20
The requisition indicated by the input ambiguous expression is calculated, and it is corrected as necessary according to the interactive processing to determine the search condition, and the recruitment information database 23 is searched.
Then, when the user corrects the wage indicated by the calculated vague expression, as shown in FIG. 11, the modified wage and the vague expression / average wage / standard deviation used in calculating the wage as the correction source Are stored in the logging file 27.

When the learning program 22 is started upon receiving the data stored in the logging file 27, as shown in the processing flow of FIG.
When one ambiguous expression stored in the logging file 27 is selected, and it is determined in subsequent step 2 whether or not this selection has been made, and when it is determined that the selection cannot be made, that is,
When all the ambiguous expressions have been selected, the process ends.

On the other hand, when it is determined in step 2 that the ambiguous expression has been selected, the process proceeds to step 3 in which the storage data having the selected ambiguous expression is selected from the unprocessed storage data of the logging file 27. In the following step 4, it is determined whether or not this selection has been made. When it is determined that the selection has been made, the process proceeds to step 5,
From the logging file 27, the corrected wage / average wage / standard deviation of the selected storage data is read.

Subsequently, in step 6, using the read corrected wage / average wage / standard deviation, the coefficient value is calculated according to a calculation formula of coefficient value = [(corrected wage) − (average wage)] / (standard deviation). Calculation, then step 7
Then, the calculated coefficient value is cumulatively added to the coefficient values calculated so far, and then the process returns to step 3.

By repeating the processing of steps 3 to 7 in this way, the cumulative addition value of the coefficient value for the ambiguous expression selected in step 1 is calculated. When it is determined that the processing of all the stored data having the ambiguous expression stored in is completed, the process proceeds to step 8, where the cumulative addition value of the calculated coefficient values is divided by the number of data having the ambiguous expression. Then, in step 9, the coefficient value of the fuzzy expression managed in the fuzzy management table 26 is changed to that of the obtained mean value, and the process returns to step 1.

As described above, the learning program 22 learns the coefficient values managed in the ambiguity management table 26 to the ones indicated by the ambiguity expression considered by the user, thereby obtaining the coefficient values managed in the ambiguity management table 26. Is changed to an appropriate one.

As a result, the search program 20 can calculate an accurate wage represented by the ambiguous expression according to the user's sense. Although the present invention has been disclosed in accordance with the illustrated embodiment, the present invention is not limited thereto. For example, in the embodiment, when an ambiguous expression is input, an ambiguous item value indicated by the ambiguous expression is calculated. However, the present invention is not limited to this. When generating the sub recruitment statistical information database 25, a configuration may be adopted in which the ambiguous item values indicated by all the ambiguous expressions are calculated in advance.

In the embodiment, the logging file 27
Then, the learning program 22 uses the data stored in the logging file 27 to obtain a coefficient value = [(corrected wage) − (average wage). ] / (Standard deviation), a coefficient value is calculated according to the formula, and the coefficient value is learned by cumulatively adding the calculated value to obtain an average value. However, the present invention is not limited to this. Instead, the search program 20 calculates a coefficient value according to this calculation formula, stores the coefficient value in the logging file 27, and
2, the coefficient value may be learned by cumulatively adding the values to obtain an average value.

[0056]

As described above, according to the database search apparatus of the present invention, it becomes possible to specify a search condition using an ambiguous expression such as "high" or "somewhat high". Will be able to easily retrieve the data you want from the database.

Then, the ambiguous expression is learned so as to match the user's sense, so that the user can accurately retrieve desired data from the database.

[Brief description of the drawings]

FIG. 1 is a principle configuration diagram of the present invention.

FIG. 2 is an embodiment of the present invention.

FIG. 3 is a processing flow executed by a setup program.

FIG. 4 is a processing flow executed by a search program.

FIG. 5 is a processing flow executed by a search program.

FIG. 6 is a processing flow executed by a learning program.

FIG. 7 is a data structure diagram of a job information database.

FIG. 8 is an example of a job recruitment statistics information database.

FIG. 9 is an example of a search condition input screen.

FIG. 10 is an example of an ambiguity management table.

FIG. 11 is an example of a logging file.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Database search apparatus 2 Database 10 Search means 11 Statistical information database 12 Generating means 13 Secondary statistical information database 14 Creating means 15 Management means 16 Calculation means 17 Decision means 18 Learning means

Claims (7)

[Claims] 1. A database search device for searching a database, comprising: a calculating unit that calculates a value corresponding to an ambiguous expression based on a search condition other than a specified ambiguous expression. . 2. A database search device for searching a database, a statistical information database for managing statistical information of each data group of the database divided by a required item value of a search condition, and a search managed by the statistical information database. Calculating means for calculating an ambiguous item value indicated by the ambiguous expression from statistical information of the ambiguous item value of the condition and a prescribed setting value assigned to the ambiguous expression specified for the ambiguous item; A search unit that determines the search condition by using an ambiguous item value calculated by the calculation unit when a search condition is designated by specifying an ambiguous expression. 3. The database search apparatus according to claim 2, further comprising: a generation unit configured to generate a statistical information database from data of the database when setting up the database. 4. The database search apparatus according to claim 2, wherein an essential item specified by the search condition does not match the one used for generating the statistical information database and becomes a part of the required item. And a creating means for creating, from the data of the statistical information database, a sub-statistical information database that manages the statistical information of each data group of the database divided by the required item value. A database search device characterized by calculating ambiguous item values indicated by ambiguous expressions using statistical information of ambiguous item values of managed search conditions. 5. The database search device according to claim 2, wherein the determining means displays the ambiguous item value calculated by the calculating means on a display screen, and displays the ambiguous item value as necessary according to the interactive processing. A database search device characterized by correction. 6. The database search apparatus according to claim 5, wherein the setting assigned to the ambiguous expression is determined by using the correction value to be corrected by the determination means and the statistical information used to calculate the ambiguous item value of the correction source. A database search device comprising a learning unit for learning a value. 7. A program storage medium for storing a program for realizing a database search device for searching a database, wherein the statistics of each data group of the database divided by the essential item values of the search condition from the data of the database Generating means for generating a statistical information database for managing information; statistical information having an ambiguous item value of a search condition managed by the statistical information database; and setting of rules to be assigned to an ambiguous expression specified for the ambiguous item A calculating means for calculating an ambiguous item value indicated by the ambiguous expression from the value, and using a fuzzy item value calculated by the calculating means when a search condition is designated by specifying a required item value and the ambiguous expression. A program storage medium, wherein a program that realizes the determination means for determining the search condition is stored.