Hybrid Computational Intelligent Attribute Reduction System, Based on Fuzzy Entropy and Ant Colony Optimization

摘要: Attribute reduction plays a crucial role in reducing the computational complexity and therefore resource consumptions area of artificial intelligence, machine learning computing applications. Rough sets are very promising technique attribute or feature selection. Fuzzy rough set hybrids have been proven to be more effective selecting important features from available data, particularly case real-time data. There is need for global searching strategies find best possible, minimal combination features, at same time maintain originality information. This paper proposes hybrid intelligent system based on fuzzy entropy, sets, ant colony optimization, which do not depend dependency degree. Experimentation conducted several UCI universal benchmark data proves this method feasible obtaining with undisturbed improved classification accuracy when compared entropy degree-based quick reduct.