R Squared

machine-learning

Definition

$R^2$

$$R^2=1-\frac{{\sum }_{i=1}^N(y_i-{\hat{y}}_i{)}^2}{{\sum }_{i=1}^N(y_i-{\bar{y}}_i{)}^2}\in [0,1]$$

where ${\hat{y}}_i$ denotes a predicted value and $\bar{y}$ the mean of the $y$ dataset.

$R^2=1$ means “theoretically perfect” and $R^2=0$ means pretty bad.