error (regression)

The error is the deviation of the observation from the true value. For example, you might believe that all observations from a sensor are noisy versions of the true value.

In contrast the residuals are the deviation of the predicted value from the observed value.