我想知道您是否知道R中有某种方法可以对测试系数进行差异化测试。
我可以为此目的使用tsDyn的TVAR.LRtest吗？我的目标是测试系数：B下降vs B上升。
代码示例：
library(tsDyn)
data(zeroyld)
TVAR(zeroyld, lag=2, nthresh=2, thDelay=1, trim=0.1, mTh=1, plot=FALSE)
> Best unique threshold 10.653
Second best: 8.124 (conditionnal on th= 10.653 and Delay= 1 ) SSR/AIC: 164.1456
Second best: 10.653 (conditionnal on th= 8.124 and Delay= 1 ) SSR/AIC: 164.1456
Second step best thresholds 8.124 10.653 SSR: 164.1456
Model TVAR with 2 thresholds
$Bdown
Intercept short.run -1 long.run -1 short.run -2
Equation short.run 0.01496509 1.0061051 0.03855754 -0.007405335
Equation long.run -0.03467035 0.3622364 1.02090855 -0.287156749
long.run -2
Equation short.run -0.03455587
Equation long.run -0.09201102
$Bmiddle
Intercept short.run -1 long.run -1 short.run -2 long.run -2
Equation short.run -0.2169119 0.83059144 0.0127026 0.13505057 0.04939883
Equation long.run -0.1386902 -0.02142815 0.8490303 0.02085644 0.16932567
$Bup
Intercept short.run -1 long.run -1 short.run -2 long.run -2
Equation short.run 1.561918 1.1780904 -0.03076567 -0.3468208 0.06986963
Equation long.run 2.037192 0.7989221 0.81823826 -0.8562527 0.05771228
Threshold value[1] "8.124 10.653"
TVAR.LRtest(data, lag=2, mTh=1,thDelay=1:2, nboot=3, plot=FALSE, trim=0.1, test="1vs")