Skip to content

Sum of the Squared Residuals

More commonly known as the Residual Sum of Squares or RSS...

The Sum of Squared Residuals is the total, across every data point in a dataset, of the squared difference between the ground-truth target value (the actual observed outcome from your training labels) and the model’s predicted output for that input.

Keep in mind that a Residual = ObservedVal - PredictedVal

Formula

\[ \mathrm{SSR} = \sum_{i=1}^{n} \bigl(y_i - \hat{y}_i\bigr)^2 \]