R2 is a metric of the quality of fit of [[Regression Testing]]. R2 is actually R _squared_, but referred to as "R two" for some reason. It tells you the **percentage of variation in the target value that's explained by the model**. It is computed using the formula: $ R^2 = 1 - \frac{SS_{res}}{SS_{tot}} $ Where: - SS_res = the _sum of squares residual_ = the error remaining after the prediction - SS_tot = the _sum of squares total_ = the spread of target values about the mean R2 can be directly added to [[Linear Regression]]s on Excel scatterplots using the "show R2" box. # Interpretation - **R^2 = 1** says your prediction is perfect - **R^2 = 0** says your prediction is only as good as just using the average value - **R^2 < 0** says your prediction is _worse_ than just using the average **** # More ## Source - Grad School ## Related -