A Novel method of Data Stream Classification Based on Incremental Storage Tree

For the characteristics of large number, fast change, high cost of random access of data stream, this paper proposes a Bayesian classification data mining algorithm based on incremental storage tree to handle the problems. Use sliding window to process data stream and divide it into several basic units, apply Principal component analysis (PCA) to compress the data from window and produce dynamic incremental storage tree, use the second power strategy to continuously update several incremental storage tree. Then use multi-classifier integration technology combines with Bayesian classification to produce Bayesian classifier. At last, adjust the weight of classifier by cyclic test, and get high classification accuracy. Detailed simulation analysis demonstrates that the presented algorithm is of high efficiency of space and time and is more stable.