时序数据挖掘概述

ime series similarity search is always a hot research topic owing to its wide application prospects in many domains, such as DNA sequence analysis, financial data analysis, sensor network monitoring, moving object tracking and motion capture, etc. It is regarded as one of the most promising technology in the future and considerable research achievements have emerged. The paper provides an overview of research progress in the field. It proposes two research phases including general stage and deep stage, and then elaborates the basic technologies. Then, it surveys data mining of time series combined with their background, which include sequence pattern discovery, correlation analysis of time series, clustering and classification of time series, and time series outlier detecting, etc. All above mentioned technologies, the pros and cons of the techniques are discussed by comparison. Finally, we outline the problems in current research and some future research issues.