Feature Selection and Rule Extraction Based on Variable Precision Rough Set

Abstract

Abstract Main factors in feature selection is dimensionality in data and decision making on partial information handling as precision classification errors. This study will demonstrate the feature selection using the rough set method, especially in handling missing information along with an ambiguous decision system with a precision variable-based approach to finding significant features and extracting the required rules. The experiment was carried out with 12 Dermatology datasets from the UCI Machine Learning Repository consisting of 11 conditional attributes and a decision variable. The experimental results show the results of the dominant conditional feature selection of 45% of the overall existing conditional features, along with more concise rules based on selected features.