Detecting Concept Drift and Classifying Data Streams
Liming Wang · Journal of Chinese Computer Systems · 2011
It is very important to mining data streams with concept drifts for many real-time decision support systems.This paper proposed a method to estimate the Confidence Interval of the true error rate of the Up-to-Date concept to a certain model based on the statistical theory.This method could detect the concept drift under a certain probability guarantee.We apply this method and KMM algorithm to the Ensemble Framework of Classifier,and give a new algorithm for data stream classification.The experimental results in the simulation and real data streams show that the algorithm is effective.