具有ARIMA(0,1,0)误差的非线性模型的异方差检验

TESTING FOR HETEROSCEDASTICITY IN NONLINEAR MODELS WITH ARIMA (0,1,0) ERRORS

本文讨论随机误差是 ARIMA( 0 ,1 ,0 )序列的非线性回归模型的异方差检验问题 .首先导出了检验的 score统计量 ,然后利用参数的正交变换 ,得到了调整的 score统计量 .最后 ,利用氯化物数据 ( Bates &Watts,1 988)

This paper discusses the tests for heteroscedasticity of randon errors in nonlinear regression modeld with ARIMA (0,1,0) errors. Score test statistic is developed. Based on the parameter orthogonality transformation, the modified score test is also derived. The chloride data (Bates & Watts, 1988) is used to illustrate our results.