用你讲的例子,我有几问题请教:
>library(tsDyn)
> data=(lynx)
> mod=lstar(log10(lynx), m=2, mTh=c(0,1), control=list(maxit=3000))
> phi1=mod$model.specific$coefficients[1:3]
> phi2=mod$model.specific$coefficients[4:6]
> gamma=mod$model.specific$coefficients[7]
> th=mod$model.specific$coefficients[8]
> z=mod$model.specific$thVar
> G=function(y,g,th) plogis(y, th, 1/g)
> tf=G(z,gamma,th)
> plot(z,tf)
> mod
> summary(mod)
Non linear autoregressive model
LSTAR model
Coefficients:
Low regime:
const1 phi1.1 phi1.2
0.4891014 1.2465399 -0.3664328
High regime:
const2 phi2.1 phi2.2
-1.0240758 0.4232669 -0.2546088
Smoothing parameter: gamma = 11.15
Threshold
Variable: Z(t) = + (0) X(t) + (1) X(t-1)
Value: 3.339
Residuals:
Min 1Q Median 3Q Max
-0.594820 -0.107360 0.014309 0.111098 0.510342
Fit:
residuals variance = 0.03805, AIC = -357, MAPE = 5.58%
Coefficient(s):
Estimate Std. Error t value Pr(>|z|)
const1 0.489101 0.204914 2.3869 0.0169929 *
phi1.1 1.246540 0.067871 18.3663 < 2.2e-16 ***
phi1.2 -0.366433 0.104301 -3.5132 0.0004427 ***
const2 -1.024076 2.430066 -0.4214 0.6734492
phi2.1 0.423267 0.172146 2.4588 0.0139415 *
phi2.2 -0.254609 0.585416 -0.4349 0.6636207
gamma 11.153834 10.004728 1.1149 0.2649120
th 3.339199 0.092748 36.0030 < 2.2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Non-linearity test of full-order LSTAR model against full-order AR model
F = 12.446 ; p-value = 1.3815e-05
Threshold
Variable: Z(t) = + (0) X(t) + (1) X(t-1)
在结果中,有不显著的系数:phi2.2 (其P值 0.6636207);
如何去掉这个不显著的phi2.2呢?一直被这个问题纠结!
另外
我在用R包中自带的数据进行门限回归学习时;发现有些问题也要向请教。
> library(TSA)
> data(prey.eq)
> prey.tar.1=tar(y=log(prey.eq),p1=4,p2=4,d=3,a=.1,b=.9,print=TRUE)
time series included in this analysis is: log(prey.eq)
SETAR(2, 1 , 4 ) model delay = 3
estimated threshold = 4.661 from a Minimum AIC fit with thresholds
searched from the 17 percentile to the 81 percentile of all data.
The estimated threshold is the 56.6 percentile of
all data.
lower regime:
Residual Standard Error=0.2341
R-Square=0.9978
F-statistic (df=2, 28)=6355.76
p-value=0
Estimate Std.Err t-value Pr(>|t|)
intercept-log(prey.eq) 0.2621 0.3156 0.8305 0.4133
lag1-log(prey.eq) 1.0175 0.0704 14.4455 0.0000
(unbiased) RMS
0.05479
with no of data falling in the regime being
log(prey.eq) 30
(max. likelihood) RMS for each series (denominator=sample size in the regime)
log(prey.eq) 0.05114
upper regime:
Residual Standard Error=0.2676
R-Square=0.9971
F-statistic (df=5, 18)=1253.556
p-value=0
Estimate Std.Err t-value Pr(>|t|)
intercept-log(prey.eq) 4.1986 1.2841 3.2697 0.0043
lag1-log(prey.eq) 0.7081 0.2023 3.5005 0.0026
lag2-log(prey.eq) -0.3009 0.3118 -0.9648 0.3474
lag3-log(prey.eq) 0.2788 0.4063 0.6861 0.5014
lag4-log(prey.eq) -0.6113 0.2726 -2.2427 0.0377
(unbiased) RMS
0.07158
with no of data falling in the regime being
23
(max. likelihood) RMS for each series (denominator=sample size in the regime)
0.05602
Nominal AIC is 10.92
我的问题是,在高区中的滞后2,3项的回归系数在一定的显著水平下,P值不算好。那么我肯定要在滞后2,3中进行选择最好的。比如,我首先选择lag3-log的系数为零,进行回归,再来看结果。。。。。。
请问:
1、在门限回归中要不要对各区的系数进行显著性或不显著性判定(用其P值)。
2、如果应当这样作的话,那么我们又如何来限制不显著的系数为零呢?
请就这个结果和prey.tar.1=tar(y=log(prey.eq),p1=4,p2=4,d=3,a=.1,b=.9,print=TRUE)语句邦助修改。谢谢!
3、另外我试了一种限制的办法,也是不行的:
>prey.tar.1=tar(y=log(prey.eq),p1=4,p2=4,fixed=c(NA,Na,0,Na,NA),d=3,a=.1,b=.9,print=TRUE)
错误于tar(y = log(prey.eq), p1 = 4, p2 = 4, fixed = c(NA,Na,0,Na,NA), :
参数((fixed = c(NA,Na,0,Na,NA))) 没有用。
二类问题是同样的。
谢谢!


雷达卡



京公网安备 11010802022788号







