MVLSC Home · Issue Contents · Forthcoming Papers
An Intuitionistic Fuzzy-Rough Attribute Selection Using Representative Samples
Aneesh Kumar Mishra, Neelesh Kumar Jain and Ravindra Kumar Singh
Selecting relevant features is an important tool for extracting knowledge from datasets with many attributes and objects. The traditional theory of rough set is a fundamental and successful tool for dealing with vagueness and inconsistency. Combining the rough set with the fuzzy set handles the information loss problem arising from the discretisation process. Still, it fails to consider the hesitancy part of any information system. A generalisation of fuzzy set known as an intuitionistic fuzzy (IF) set has more real-world applications to confront uncertainty and ambiguity than the fuzzy set. So, the combination of rough set and IF set not only deals with vagueness but also able to consider the hesitancy available in any real-world data. In this work, we propose an IF rough set model based on representative samples and its application in the area of attribute reduction of high-dimensional datasets. First, we defined the representative sample-based intuitionistic fuzzy rough set and then presented an algorithm to calculate the reduction of a dataset using the degree of dependency method. Mathematical theorems are applied to validate the presented model theoretically. Experimental analysis is also discussed to validate the proposed technique. Finally, we applied our proposed method to improve the prediction of antifungal peptides.
Keywords: Attribute reduction, representative sample, rough set, intuitionistic fuzzy set, intuitionistic fuzzy rough set, decision making