0

I have different models (linear (lm), gls, GARCH) where I would like to check them for Heteroscedasticity.

However for the lm-model it is very easy, visually and with tests as follows:

fit1 <- lm(formula = X0~X1 + X5 + X7 + X8 + X9 + X10 + X11 + X12, 
          data = my_data, weights = NULL)
    residualPlots(fit1)
    bptest(fit1)
    ncvTest(fit1)

For the other models it is not so easy! Do you have any ideas?

Ιndicatively reported, that I have a sample with 15 variables of X and one dependent variable Y.

fit22<-gls(Y~X1+X2+X3+X4+X5+X6+X7+X8+X9+X10+X11+X12+X13+X14+X15,data=my_data,correlation = corARMA(p=0,q=1,form = ~ 1))
Marz
  • 15
  • 6
  • could you extract the residuals from your model objects and pass them to a function like `MASS::boxcox()`? – Nate Jan 17 '17 at 22:41
  • Are you sure that your other models have the same assumptions regarding heteroskedasticity? – IRTFM Jan 19 '17 at 04:54
  • Maybe they don't have the same assumptions, but if I want to check them for heteroskedasticity, what can I do? – Marz Jan 19 '17 at 10:17

0 Answers0