Rule Discovery in Databases based on a Simultaneous Confidence Evaluation

13 References
  • SUZUKI Einoshin
    Division of Electrical and Computer Engineering, Yokohama National University

Bibliographic Information

Other Title
  • 信頼性同時評価に基づく確率的 if-then ルール発見手法

Search this article

Description

This paper presents an algorithm for discovering probabilistic if-then rules, which represent an important subclass of production rules. Previous approaches have either ignored the reliability evaluation or only evaluated the reliability of the generality. Consequently, they tend to discover a huge number of rules, some of which are unreliable in their accuracies. In order to circumvent these difficulties we simultaneously evaluate the reliability of the generality and the accuracy based on the normal approximations of the multinomial distributions. A novel pruning method is employed for improving its time efficiency. Our approach has been validated experimentally using 21 benchmark data sets.

Journal

  • IPSJ SIG Notes

    IPSJ SIG Notes 48 53-60, 1997-11-13

    Information Processing Society of Japan (IPSJ)

References(13)*help

See more

Details 詳細情報について

  • CRID
    1573668926989709440
  • NII Article ID
    110002934069
  • NII Book ID
    AN10114171
  • Text Lang
    ja
  • Data Source
    • CiNii Articles

Report a problem

Back to top