1 2 points Residual analysis is useful for all but which of the following: checking for unequal variances testing to make sure the interactions belong in our best model looking for outliers helping make sure we found the best model 2 3 2 points It is not a problem if you have correlated predictor variables, but it is a problem if you have correlated regression errors. True False 2 points Holt's smoothing method only accounts for long-term trend. True False 2 points What is unique about time series forecasting: the predictor variable is a future time period it is generally done to avoid violating assumptions it almost invariably involves economic data the experimental unit is a unit of time MacBook Air