Co-anomaly Event Detection in Multiple Temperature Series

摘要

Co-anomaly event is one of the most significant climate phenomena characterized by the co-occurrent similar abnormal patterns appearing in different temperature series. Indeed, these co-anomaly events play an important role in understanding the abnormal behaviors and natural disasters in climate research. However, to the best of our knowledge the problem of automatically detecting co-anomaly events in climate is still under-addressed due to the unique characteristics of temperature series data. To that end, in this paper we propose a novel framework Sevent for automatic detection of co-anomaly climate events in multiple temperature series. Specifically, we propose to first map the original temperature series to symbolic representations. Then, we detect the co-anomaly patterns by statistical tests and finally generate the co-anomaly events that span different sub-dimensions and subsequences of multiple temperature series. We evaluate our detection framework on a real-world data set which contains rich temperature series collected by 97 weather stations over 11 years in Hunan province, China. The experimental results clearly demonstrate the effectiveness of Sevent.