Anomaly Detection Using Probability Density Space Partitioning-based Symbolic Time Series Analysis

Early detection of anomalies in machine system is essential for prevention of production accident.Among various methods,symbolic time series has been widely used.The effectiveness of this method is highly dependent on the procedure of symbol sequence generation.This paper presents a novel partitioning method called probability density space partitioning for the symbol sequence generation.In this partitioning approach,a time series is divided into several equal-sized regions based on the probability density distribution and each region is represented by a symbol.To verify the effectiveness of probability density space partitioning-based symbolic time series analysis,bearing test-to-failure experiments are conducted.The experimental results indicate that probability density-based method is more sensitive to the anomaly in the system than traditional partition based methods.
anomaly detection symbolic time series analysis symbol sequence generation probability density
Shijie Hu Yuning Qia Ruqiang Yan
School of Instrument Science and Engineering Southeast University Nanjing,China
国际会议
南京
英文
1-4
2013-08-20(万方平台首次上网日期,不代表论文的发表时间)