Active Fuzzy Based Learning for Incomplete Decision Support System

摘要：

In this paper we explore active learning concept using fuzzy based approach to construct highly interpretable and a good accuracy in incomplete decision support system. The problem occurs when there is significantly less information or incomplete observations of the target concept. This incomplete information introduces uncertainty into decision modeling evaluation. We propose efficient active learning to select informative instances from small pool of expert knowledge and measure the degree of uncertainty of rank objects during decision modeling for generating simple and comprehensible decision rule sets. In planting material selection, our experimental results show that the proposed method can improve classification significantly, compared to classifier trained with a large datasets.

关键词： active learning decision tree classification uncertainty

作者: Mohd Najib Mohd Salleh Rozaida Ghazali Muhaini Othman

作者单位: Faculty of Computer Science and Information Technology Universiti Tun Hussein Onn Malaysia,Parit Raja 86400 Batu Pahat, Johor, Malaysia

会议类型: 国际会议

会议名称: 2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)

会议地点: 上海

会议语种:英文

页码: 427-430

在线出版日期: 2010-12-25（万方平台首次上网日期，不代表论文的发表时间）