By combining non-parameter Z test method for unit root test of time-series with IPS test method for panel data unit root test, we propose a new non-parameter unit root test method for panel data.This method solves the...By combining non-parameter Z test method for unit root test of time-series with IPS test method for panel data unit root test, we propose a new non-parameter unit root test method for panel data.This method solves the unit root test’s problem that {ε_ it } is L-order auto-correlaiton.Using random simulation method,we compare LL test with non-parameter unit root test for panel data.We found that the non-parameter unit root test is superior to LL test in this case.展开更多
本文引入Fisher-Yates提出的计分函数和Van der Waerden提出的计分函数,对只有一个变点的位置参数模型的假设检验问题分别给出了四个检验统计量.利用Monte-Carlo随机模拟的方法,求出了检验的渐近临界值,并且对本文提出的检验,以及Pettit...本文引入Fisher-Yates提出的计分函数和Van der Waerden提出的计分函数,对只有一个变点的位置参数模型的假设检验问题分别给出了四个检验统计量.利用Monte-Carlo随机模拟的方法,求出了检验的渐近临界值,并且对本文提出的检验,以及Pettitt在文[1] 中提出的检验,Schechtman和Wolfe在文[2]中提出的检验的势进行了比较。展开更多
文摘By combining non-parameter Z test method for unit root test of time-series with IPS test method for panel data unit root test, we propose a new non-parameter unit root test method for panel data.This method solves the unit root test’s problem that {ε_ it } is L-order auto-correlaiton.Using random simulation method,we compare LL test with non-parameter unit root test for panel data.We found that the non-parameter unit root test is superior to LL test in this case.
文摘本文引入Fisher-Yates提出的计分函数和Van der Waerden提出的计分函数,对只有一个变点的位置参数模型的假设检验问题分别给出了四个检验统计量.利用Monte-Carlo随机模拟的方法,求出了检验的渐近临界值,并且对本文提出的检验,以及Pettitt在文[1] 中提出的检验,Schechtman和Wolfe在文[2]中提出的检验的势进行了比较。
