The previous chapter has introduced several learning models with particular focus on linear parametric form. In this chapter, these learning models will be utilized in a score metric to form the learning objective function. The score metric is also known as the loss function which measures the error of each learning instant. The sum of the learning scores or error losses over a set of training samples forms the learning objective function, which is also known as the learning criterion function. Three basic types of learning objective or criterion function, namely, the regression accuracy, the classification accuracy, and the ranking and operating characteristics will be presented in this chapter.