As we introduce new variables, there is a point beyond which the performance of statistical learning tends to degrade. That is, the gain from useful information contained in any new variable no longer outweighs the increase in data sparsity. Given the following three situations, please discuss what kind of training results we will have. Consider p n (no. predictors no. observations) Consider p n (no. predictors no. observations) Consider p n (no. predictors no. observations)