Skip to Main Content
It is obvious that internet has become a key media to share resources and exchange information. As a special category of social activities, the behavior from network users normally shows its complexity and diversity, which makes people pay an increased attention to study and manage it. Based upon the formation mechanism of ant colony, this paper proposes an ant colony algorithm to do cluster analysis of network user behavior. The advantages of this algorithm include: zero user interference; no predefined cluster numbers; arbitrary-shaped cluster recognition; no requirement for date types; outlier points insensitivity and so on. Our experimental analysis shows the effectiveness of this algorithm.