Mining Concept Drifts from Data Streams Based on Multi-Classifiers
Yue Sun, Guojun Mao, Xu Liu, Chunnian Liu · 2007
Mining concept drifts is one of the most important fields in mining data streams. In this paper, a new ensemble algorithm called ICEA is proposed for mining concept drifts from data streams, which uses ensemble multi-classifiers to detect concept changes from the data streams in an incremental way. The experimental results show that ICEA algorithm performs higher accuracy and better adaptability than the popular methods such as SEA algorithm.