ABSTRACT
In the previous chapter we discussed discrepancy measures for checking the fit of model to data. In this chapter we seek not to check models but to compare them and explore directions for improvement. Even if all of the models being considered have mismatches with the data, it can be informative to evaluate their predictive accuracy and consider where to go next. The challenge we focus on here is the estimation of the predictive model accuracy, correcting for the bias inherent in evaluating a model’s predictions of the data that were used to fit it.
