A Taxation attribute reduction based on genetic algorithm and rough set theory

Abstract

Selection of taxation attributes is one difficult question in analyzing the sources of taxation. This paper introduces genetic-algorithm-based rough set attribute reduction algorithm into the job of taxation attribute reduction. By referring to the concept of dependability in rough set, this method optimizes the configuration of fitness function, improves the convergence of original algorithm and changes the limitation of current attribute reduction in genetic algorithm. This algorithm fundamentally realizes the selection of comparatively small attribute sets with the presupposition that the data classification ability is not changed. It is valid after being tested.