So far, the focus has been on nonparametric and intermediate tranche model classes, clustering, and dimension reduction. So, the modeling techniques presented have been abstract and general. The perspective has been to search for a reasonable model class via unsupervised learning or dimension reduction or to assume a reasonable model class had been identified. In both cases, the goal was understanding the model class so the problem could become finding elements of the class that fit reasonably well and gave good predictions. The focus was on the model class as a whole more than on the models in the class. In this chapter, the focus is on the models themselves rather than the general properties of the class they came from.
KeywordsAkaike Information Criterion Bayesian Information Criterion Inclusion Probability Ordinary Little Square Estimator Mean Square Prediction Error
Unable to display preview. Download preview PDF.