An Attribute Reduction Method Based on Ant Colony Optimization
Yuanchun Jiang, Yezheng Liu · 2006
Attribute reduction is an important process in data mining based on rough set theory. Regarding the significance of attribute defined from the viewpoint of information theory as heuristic information and introducing it into Ant Colony Optimization (ACO), an effective heuristic ACO method is proposed to search the minimal relative reduction. Firstly, we research the model of attribute reduction and analyze the differences between TSP and attribute reduction. Secondly, we redefine the heuristic information and the pheromone updating rule. Lastly, the formation process of solution is researched. Experiments show that the proposed method can reduce attributes effectively.