科研之友
微信
新浪微博
Facebook
分享链接
摘要
Clustering Analysis is one kind of pattern recognition that not to be supervised. The Clustering algorithm based on object function resolves the clustering problem into optimization problem, thereby it becomes to the main investigatory stream nowadays. But it has some shortcomings such as its sensitivity to initial condition, and it is easy to fall in local peak. To overcome these deficiencies, ant colony optimization algorithm is applied to clustering analysis and a novel clustering based on an improved ant colony optimization algorithm is proposed. Theoretical analysis and experiments show this method is faster and more efficient to convergence upon the optimal value in the whole field.
