SWSD: An Abnormal Detection Algorithm on Unequally Spaced Time Series for Disaster Prediction
Qinyong Li, Haoming Guo, Xinjian Shan, Zhixin WANG, Yanyan Wei, Jianxiu Bai, Qiuhong ZHANG · 2020
Natural disasters cause enormous damage. Time series analysis is an effective way to predict natural disasters. To cope with classical methods can not directly applied to unequally spaced series, in this paper we propose a series of methods for abnormal detection on unequally spaced time series for disaster prediction. The main contribution of this paper has three-folds: (1) we define the concept of optimal interval for unequally spaced series; (2) based on the concept of optimal interval, we propose an algorithm to fill missing data in unequally spaced series; and (3) we propose an algorithm, called SWSD, which can be used in unequally spaced time series abnormal detection for disaster prediction. Experiment on real observation data sets demonstrates the effectiveness of the proposed work.