Method and apparatus of supporting creation of classification rules

• US 8,176,050 B2
• Filed: 11/17/2005
• Issued: 05/08/2012
• Est. Priority Date: 08/19/2005
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Inakoshi, Hiroya, Uchino, Kanji, Inoue, Daigo, Hanno, Hirokazu
• Primary Examiner(s): Lie, Angela M
• Assistant Examiner(s): MACKES, KRIS E
• Application Number:
  US11/281,956
• Publication Number:
  US 20070043690A1
• Time in Patent Office: 2,364 Days
• Field of Search: 706/20, 707/737, 707/793
• US Class Current: 707/737
• CPC Class Codes: G06F 16/90 Details of database functions independent of the retrieved data types

Abstract

This invention is to appropriately support creation of a classification rule for a new data item. This method includes: storing a new data item and a category of the new data item into a storage; extracting feature patterns each including a condition including a feature element of the new data item stored in the storage and a corresponding category from data stored in a correct answer data storage, which stores data items and categories of the data items, and storing the feature patterns into a feature pattern storage; and grouping the feature patterns stored in the feature pattern storage into a first set whose feature pattern matches with the category of the new data item stored in the storage and a second set whose feature pattern does not match with the category of the new data item, and storing a grouping result into a group data storage. Accordingly, the feature pattern, which is a candidate of the classification rule, is automatically generated, and furthermore, it becomes possible to select a specific feature pattern as a classification rule to be applied hereinafter from the first set.

Claims

1. A method for supporting creation of a classification rule, comprising:

first extracting, by a processor of a computer that accesses a correct answer data storage unit storing correct answers each including a data item and a corresponding first correct category of the data item, and a classification rule storage unit storing classification rules each including a condition part that includes at least one keyword and a corresponding second category, a plurality of feature patterns from the correct answer data storage unit, wherein each of the plurality of feature patterns includes at least one feature element included in a new data item inputted by a user and a first correct category corresponding to a data item including the at least one feature element;

second extracting, by the processor, one or plural classification rules whose condition part satisfies the new data item from the classification rule storage unit;

grouping, by the processor, the plurality of feature patterns into a first group of feature patterns that are satisfied with a condition part of one classification rule included in the one or plural classification rules extracted in the second extracting and have a first correct category different from a second category of the one classification rule, a second group of feature patterns that are satisfied with the condition part of the one classification rule and have a first correct category identical to the second category of the one classification rule, and a third group of feature patterns that are not satisfied with the condition part of each of the one or plural classification rules extracted in the second extracting; and

outputting, by the processor, data of the first group in association with the one classification rule, data of the second group in association with the one classification rule and data of the third group in association with data representing that no condition parts of the one or plural classification rules satisfy feature patterns included in the third group.

2. The method as set forth in claim 1, further comprising:

registering, by the processor, a selected feature pattern of the feature patterns included in the first group and a second feature pattern that is included in the second group and is satisfied with a condition part of a classification rule, which satisfies the selected feature pattern, into the classification rule storage unit; and

removing, by the processor, the classification rule whose condition part satisfies the selected feature pattern and the second feature pattern from the classification rule storage unit.

3. The method as set forth in claim 2, further comprising:

registering, by the processor, a feature pattern selected by a user among the feature patterns included in the third group into the classification rule storage unit.

4. The method as set forth in claim 1, further comprising:

calculating, by the processor, a score of each of the feature patterns extracted in the second extracting based on feature elements of each of the feature patterns extracted in the second extracting; and

registering, by the processor, a feature pattern extracted in the second extracting one by one into the classification rule storage unit as a new classification rule in descending order of the score of the feature pattern, until an error rate, which changes according to the number of occurrences that data items stored in the correct answer data storage unit are misclassified, increases.

5. The method as set forth in claim 1, further comprising:

detecting, by the processor, that the number of feature patterns included in the first group satisfies a condition concerning the number of feature patterns included in the second group; and

upon the detecting, changing, by the processor, a condition of the feature patterns included in the first group to a condition including a negative rule.

6. The method as set forth in claim 1, further comprising:

first detecting, by the processor, that the number of feature patterns included in the first group satisfies a condition concerning the number of feature patterns included in the second group;

upon the first detecting, changing, by the processor, a condition of the feature patterns included in the first group to a condition including a negative rule;

second detecting, by the processor, that the number of feature patterns included in the second group satisfies a condition concerning the number of feature patterns included in the first group; and

upon the second detecting, changing, by the processor, a condition of the feature patterns included in the second group to a condition including a negative rule.

7. The method as set forth in claim 1, further comprising:

presenting, by the processor, the user with the feature patterns included in the first group;

accepting, by the processor, selection of a feature pattern from the user;

registering the selected feature pattern into the classification rule storage unit;

identifying, by the processor, a second feature pattern that is included in the second group and is satisfied with a condition part of a classification rule, which satisfies the selected feature pattern,

registering, by the processor, the second feature pattern into the classification rule storage unit; and

removing, by the processor, the classification rule whose condition part satisfies the selected feature pattern and the second feature pattern from the classification rule storage unit.

8. The method as set forth in claim 1, further comprising:

removing a specific classification rule whose condition part satisfies the feature pattern included in the first group and in the second group from the classification rule storage unit;

registering the feature patterns that are included in the second group and are associated with the specific classification rule into the classification rule storage unit;

presenting the user with the feature patterns included in the first group and the third group;

accepting selection of a feature pattern from the user; and

registering the selected feature pattern into the classification rule storage unit.

9. A computer-readable, non-transitory storage medium storing a program for causing a computer to execute a procedure comprising:

first extracting, from a correct answer data storage unit storing correct answers each including a data item and a corresponding first correct category of the data item, a plurality of feature patterns from the correct answer data storage unit, wherein each of the plurality of feature patterns includes at least one feature element included in a new data item inputted by a user and a first correct category corresponding to a data item including the at least one feature element;

second extracting, from a classification rule storage unit storing classification rules each including a condition part that includes at least one keyword and a corresponding second category, one or plural classification rules whose condition part satisfies the new data item;

grouping the plurality of feature patterns into a first group of feature patterns that are satisfied with a condition part of one classification rule included in the one or plural classification rules extracted in the second extracting and have a first correct category different from a second category of the one classification rule, a second group of feature patterns that are satisfied with the condition part of the one classification rule and have a first correct category identical to the second category of the one classification rule, and a third group of feature patterns that are not satisfied with the condition part of each of the one or plural classification rules extracted in the second extracting; and

outputting data of the first group in association with the one classification rule, data of the second group in association with the one classification rule and data of the third group in association with data representing that no condition parts of the one or plural classification rules satisfy feature patterns included in the third group.

10. An apparatus for supporting creation of a classification rule, comprising:

a correct answer data storage unit storing correct answers each including a data item and a corresponding first correct category of the data item;

a classification rule storage unit storing classification rules each including a condition part that includes at least one keyword and a corresponding second category; and

a processor configured to execute a procedure comprising;

first extracting, from the correct answer data storage unit, a plurality of feature patterns from the correct answer data storage unit, wherein each of the plurality of feature patterns includes at least one feature element included in a new data item inputted by a user and a first correct category corresponding to a data item including the at least one feature element;

second extracting, from the classification rule data storage unit, one or plural classification rules whose condition part satisfies the new data item;

grouping the plurality of feature patterns into a first group of feature patterns that are satisfied with a condition part of one classification rule included in the one or plural classification rules extracted in the second extracting and have a first correct category different from a second category of the one classification rule, a second group of feature patterns that are satisfied with the condition part of the one classification rule and have a first correct category identical to the second category of the one classification rule, and a third group of feature patterns that are not satisfied with the condition part of each of the one or plural classification rules extracted in the second extracting; and

outputting data of the first group in association with the one classification rule, data of the second group in association with the one classification rule and data of the third group in association with data representing that no condition parts of the one or plural classification rules satisfy feature patterns included in the third group.