Reduction of low frequency concepts in the concept drift based on the frequency
Xuegang Hu · Journal of Hefei University of Technology · 2009
The traditional algorithms of classification are no longer suitable because of the effect of the concept drift in data streaming.The research on fast and stable data mining systems with high accuracy has become more valuable in both theory and practice.The effect of low frequency concepts on time and space performance is studied and an LFCR algorithm is proposed in this paper,which reduces the low frequency concept with some mechanisms.The extensive experimental study shows that LFCR has improved performance in classification in comparison with RePro.