Review of Segmenting Algorithms for Streaming Time Series
(0) With the development of science and technology, a new problem facing data mining community is to discover knowledge from data with complicated type. In recent years, mining data stream has attracted many researchers interest, an important scenario is streaming time series. The representation of data is the key to discover knowledge. The segmentation problem is defined, including formalization definition. The authors focus on the segmenting algorithms for streaming time series and categorize the proposed approaches into four types, namely, improved methods based on static representation, Sliding Window based methods, combined methods based on static representation and Sliding Window, prediction mechanism. Each category is discussed in detail. Especially, the features of four types are analyzed and compared. Finally, future research directions about segmenting algorithms for streaming time series are outlined.
Guiling Li Yuanzhen Wang Min Li
College of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan, 430074, China School of Computer Science, China University of Geosciences, Wuhan, 430074, China
国际会议
武汉
英文
212-216
2008-12-19（万方平台首次上网日期，不代表论文的发表时间）