Application of Statistical Control Charts to detect unusual frequency of earthquake in the world

Fariha Taskin ； Mohammad Shahed Masud

《Journal of Data Science》 18卷1期 (2020/01) Pp. 44-55

https://doi.org/10.6339/JDS.202001_18(1).0002

摘要

Earthquake in recent years has increased tremendously. This paper outlines an evaluation of Cumulative Sum (CUSUM) and Exponentially Weighted Moving Average (EWMA) charting technique to determine if the frequency of earthquake in the world is unusual. The frequency of earthquake in the world is considered from the period 1973 to 2016. As our data is auto correlated we cannot use the regular control chart like Shewhart control chart to detect unusual earthquake frequency. An approach that has proved useful in dealing with auto correlated data is to directly model time series model such as Autoregressive Integrated Moving Average (ARIMA), and apply control charts to the residuals. The EWMA control chart and the CUSUM control chart have detected unusual frequencies of earthquake in the year 2012 and 2013 which are state of statistically out of control.

關鍵字

CUSUM chart ； EWMA chart ； ARIMA