In [[Machine Learning]], the error of an observation is the deviation of the observed value from the rue value of the quantity of interest (the target). This deviation can be measured using different functions, which are **error metrics**.