QUANTITATIVE EXTRACTING METHOD AND EXAMPLE FOR TREND TRANSITION POINTS OF CRUSTAL DEFORMATION OBSERVATION DATA
Liu Guan-zhon · Inland Earthquake · 2013
Trend transition is important information in crustal deformation data analysis. In this paper,for extracting the transition points of time series quantitatively and precisely,we study the improved sequential cluster method with moving time-window and compile the computer program. As an example,we deal with the ground tilt data about NS in Xinyuan seismic station from 1992 to 2006with 4 years window-length and 2 years step-size. The results show that:( 1) the method of improved sequential cluster analysis has an explicit statistic meaning,and it can reduce the subjectivity and arbitrariness in study.( 2) The algorithm of this method is very simple,can use easily in analysis and prediction for crustal deformation observation.( 3) The calculation result of Xinyuan station suggest that the trend transition was frequent and obvious relatively before 2001,whereas,after2001,it was rare and more like linear.