Integrating fractal dimensionality reduction with cluster evolution tracking

Detecting and tracking of cluster evolution has always been crucial to the stream data mining. While it, in high dimensional stream data environment, becomes more difficult under the interaction between dimensionality reduction and cluster evolution condition. The past has been focus on cluster evolution occurred in the reduced dimensionality space. Dimensionality reduction before the cluster evolution option, however, can not cope with the abrupt changes which are common in stream data. There is the demand of the dimensionality reduction during the process of the cluster evolution, which is the most popular case. In the paper, we pay more attention for the interaction between dimensionality reduction and cluster evolution in the inconstant high dimensional stream data. And on this basis, we propose the adaptive cluster evolution tracking algorithm which integrated the on-line fractal dimensionality reduction technique. Experimental results over a number of real and synthetic data sets show that the method proposed are both effectiveness and efficiency.