Incremental Web User Clustering Based on Ants´ Chemical Recognition System Algorithm

This paper introduces an efficient incremental clustering method based on Ants´ Chemical Recognition System Algorithm (ACRSA). We apply it to Web user clustering and compare the accuracy rate of personalized recommendation with ACRSA. This paper also introduces a new cluster-dissolution mechanism into ACRSA to make the result more natural. The experimental results show that this method can achieve incremental clustering adaptively and efficiently, and give more accurate recommendations than ACRSA.