Statistical inference of geographically and temporally weighted regression model

The fundamental issues of statistical inference related to geographically and temporally weighted regression (gtwr) model are studied. initially,the test statistics for hypothesis testing problems of global stationarity,spatial nonstationarity and temporal nonstationarity are proposed by analysis of variance technique. the heteroscedasticity in gtwr model is detected and score test statistic is provided. finally,an approximation method is proposed to compute the p-values for aforementioned test statistics. a simulation study is carried out to assess the performance of these test methods,and a real example of per capita gdp in chinese 92 cities is given. © 2015 pakistan journal of statistics.