Study on the Classification Algorithm for Streaming Data with Concept Drift
Hou Chuan-yu · Journal of Suzhou University · 2008
The traditional classification algorithms have no longer fit for the streaming data,because of the streaming data may grow without limit,arriving in sequence with uncertain speed,and so on.It is difficult to design a efficient and high accuracy classification system for the streaming data for it's uncertain changing.So research the fast,high accuracy and stable classification system to be high worth in both abstract and applying,since the data stream existing in many area.In recent years a great deal of research work were be done to solve the classification problem efficiently and accurately about streaming data,which have underlying concept drift.In this paper some classification algorithms,which suitable for resolving the problem of underlying concept drift in streaming data are introduced.By analyzing them,it can provide help to researchers in this domain.