Lecture 6

It is often useful to have a simple summary of how much variation in the data that our fitted model has actually explained. Any model with at least one explanatory variable will typically be better than fitting a model with only an intercept term with no explanatory variables. In practice the most reduced and smallest model is usually considered to be the intercept only model and we typically compare our fitted model with at least one explanatory variable to a model that fits an overall mean for the response data, i.e. the intercept only model.