I estimated autoregressive model on a time series varibles, by selecting the lag length based on correlogram.
what should be desired properties of the residuals derived from autoregressive model.
is it essential that the residuals should have the properties like
1. normal distribution
2. free from autocorrelation
3. homoscedasticity
4. no multicollinearity between the autoregressive terms used in the model as independent variables.
thank you Prof. Ette Etuk. for your answer
what is the remedial measure if the residuals are not normally distribued.
thanks in advance.
thank you professor Ette Etuk for your prompt response.
i have one more doubt. uncorrelatedness of residuals means that residuals should not be correlated with the independent variables in the model.am I right?
please clarify my doubt.
Residuals in autoregressive models are serially autocorrelated by model specification, as a result the most desirable properties of these family of models is iid and normality distribution of the residuals. Autocorrelation does not refer to the correlation of the independent variables with the residuals.
dear Prof. Ibrahim Ahmed Onour
if residuals are not normally distributed what is the remedial solution to the problem