地震前后基准站GPS数据异常提取算法.

Most research studies on forecasting of earthquakes focus on thermal or electrical aspects. Scholars have proved GPS time series also contains precursory information for large earthquakes. For the problems mentioned before, three typical earthquakes hit America from 2001 to 2010 were investigated by using GPS data of several GPS reference stations near the epicenters. This paper proposed an algorithm based on Martingale theory to detect the anomalies in GPS data. The experimental results show that the proposed algorithm can effectively reflect the change process in GPS data before and after large earthquakes, which offers more possibilities for earthquake forecast from GPS data. [ABSTRACT FROM AUTHOR]