A Research of machine learning of object oriented analysis knowledge by induvtive reasoning
| Project/Area Number |
09680377
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Aoyama Gakuin University |
|---|
Principal Investigator |
HARADA Minoru College of Science and Engineering, Aoyama Gakuin University, Associate Professor, 理工学部, 助教授 (10218654)
|
|---|
| Project Period (FY) |
1997 – 1999
|
| Project Status |
Completed (Fiscal Year 1999)
|
| Budget Amount *help |
¥3,000,000 (Direct Cost: ¥3,000,000)
Fiscal Year 1999: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1998: ¥1,100,000 (Direct Cost: ¥1,100,000)
Fiscal Year 1997: ¥1,100,000 (Direct Cost: ¥1,100,000)
|
|---|
| Keywords | theory revision / discrimination rule / progol / prolog / object oriented analysis automation / rule base maintenance / 帰納学習 / progol / 帰納推論 / 格フレーム生成 / オブジェクト指向分析 / 自然語処理 |
|---|
| Research Abstract |
A lot of rule-based systems are presently in use. However, the reliabilities of these systems depend primarily on the rules initially given by the expert on the domain. No expert is capable of preparing a perfect rule base because creating a complete set of rules would require that every possible situation has been considered, Hence, if an incorrect execution result comes out, it is usual to correct the rule base by manual operation. The purpose of the present study is to develop system THERES (Theory Revision System) which automates this correction of rules. The rule base which THERES revises is a Horn clause set described with prolog. The predicates which compose the rule base are assumed to be divided into the correct predicates (called basic predicates) which need not be revised and the predicates (called rule predicates) as the target of maintenance, whose body section is described by basic predicates (single rule decidable rule base). After the rule base system is applied to a ne
… More
w problem, only by dividing the results into the positive cases and the negative cases and also by pointing out the new positive cases which were not derived by the rule base but are thought to be correct, THERES automatically revises the rule base so as to solve this problem correctly. THERES first searches incorrect rule R, then acquires discrimination rule D derived from R using the inductive learning system Progol to separate positive and negative examples of R, and finally adds the body of D to that of R in the form of a conjunction. The proposed system corrects incorrect rules as if adding missing regulations on exceptions to the existing rules created by an expert. Therefore, it is more practical than to replace the existing rules by the rules acquired by inductive learning from the cases. The present study applies THERES to revise the rule base of an automatic object-oriented analysis system CAMEO, a practical rule base system previously developed by us for automating object oriented analysis, in order to discuss the effectiveness of the proposed method. Less
|
Report
(4 results)
Research Products
(12 results)
