An efficient clustering approach using ant colony algorithm in mutidimensional search space

Clustering is an important data analysis technique and it widely used in many field such as data mining, machine learning and pattern recognition. Ant colony optimization clustering is one of the popular partition algorithm. However, in mutidimensional search space, its results is usually ordinary as the disturbing of redundant information. To address the problem, this paper presents MD-ACO clustering algorithm which improves the ant structure to implement attribute reduction. Four real data sets from UCI machine learning repository are used to evaluate MD-ACO with ACO. The results show that MD-ACO is more competitive.