residual

The residual is the distance between \(\tilde{y}\) (what the model predicted) and the actual value \(y\).

It is particularly important for the Linear Regression model, that minimizes residual sum of squares (RSS) to estimate weights.