Homoscedasticity and Autocorrelation


Ambassador to the humans
I don't see any issues with unequal variance. To assess the autocorrelation you would want to plot the residuals in time order. Maybe you did that but...I don't read whatever language that is so I can't tell


New Member
Thank you for advice. I think that i don´t want to have the residuals in time order. I thought it is requiered for the multiple regression that the residuals are independent, isn´t it?


Fortran must die
There are a range of formal test of autocorrelation. I would use one of those rather than looking at the residuals. I think Breusch-Godfrey is the best test of this although its been a while.