The Self-Adaptive Integrated Algorithm On Time-Sensitive Cluster Evolution Tracking

There are many works on the stream evolution tracking methods and dimensionality reduction techniques respectively during the last decades, however, till now, the works focus on the interaction scheme between the dimensionality reduction and the cluster evolution, and studies on how to use this interaction to enhance the positive feedback between the two operations are rare. To this end, we transform the dimensionality reduction problem into a biobjective optimisation with the maximal fractal dimension and the minimum attribute number restriction simultaneously, and propose heuristic rules to resolve the bi-objective optimisation problem. Furthermore, we discuss the interaction between the dimensionality reduction operation and the cluster evolution in the time decayed stream data and illustrate the integration of selfadaptive sample technique with fractal cluster technique on time-sensitive cluster evolution tracking. The performance experiments over a number of real and synthetic data sets illustrate the effectiveness and efficiency provided by our approach.