ynum-xnum-use-regr-lin

(0.0) Assumptions
(0.1) If sample data used, sample is randomly selected from population
(0.2) y = linear( x )
(0.3) residuals = y_obs - y_pred 
                = evenly scattered 
                = equal variance
(0.4) r   = rho   = correl
                  = Multiple R
      r^2 = rho^2 = R Square
                  = % Data Explained by Model
(0.5) residual ~ Norm
      +-- Histogram of Residuals
      +-- BoxPlots
(1.0) Multi-collinearity   
      (1.1) Remove one variable   
      (1.2) Regularisation
            Add penalty
               Ridge Regression
               Lasso Regression