Analysis of Decision-Theoretic Rough Set Model Based on Error Rate

International Journal of Artificial Intelligence and Knowledge Discovery

Sriram G. Sanjeevi²  Srilatha Chebrolu¹ 

[1] NIT Warangal;National Institute of Technology Warangal

关键词: rough set theory;    decision-theoretic rough set model;    attribute reduction;    error rate;    genetic algorithm;   

DOI  :  

学科分类：建筑学

来源: RG Education Society

【 摘 要 】

This paper analyses decision-theoretic rough set model based on the objective of minimizing the number of objects wrongly classified into various probabilistic regions. While defining the objective function, the risk of misclassifying an object into a probabilistic region is also considered. Thus the threshold parameters are determined through optimizing this objective function. At these determined threshold parameter values a minimal attribute reduct is found through optimizing the same objective function. Experimental results on various UCI ML repository data sets are compared with the correlation based feature selection method and consistency based feature selection method by taking classification accuracy as the metric of comparison. These comparisons shows the proposed attribute reduction approach based on decision-theoretic rough set model is achieving a  higher  classification accuracy after attribute reduction. 

【 授权许可】

Unknown