Discovering an Evolutionary Classifier over a High-speed Nonstatic Stream

Abstract

With the emergence of large-volume and high-speed streaming data, mining data streams has become a focus of increasing interests. The major new challenges in streaming data mining are as follows: (1) since streams may flow in and out indefinitely and in fast speed, it is usually expected that a stream mining process can only scan a data stream once; and (2) since the characteristics of the data may evolve over time, it is desirable to incorporate evolving features of data streams. This paper investigates the issues of developing a high-speed classification method on streaming data with concept drifts. Among several popular classification techniques, the naı̈ve Bayesian classifier is chosen due to its low construction cost, easiness for incremental maintenance, and high accuracy. An efficient algorithm, called EvoClass (Evolutionary Classifier), is devised. EvoClass builds an incremental, evolutionary Bayesian classifier on streaming data. A train-and-testmethod is employed to discover the changes in the characteristics of the data and the need for construction of a new classifier. In addition, divergence is utilized to quantify the changes in the classifier and inform the user what aspect of the data characteristics has evolved. Finally, an intensive empirical study has been performed that demonstrates the effectiveness and efficiency of the EvoClass method