Data Modeling

The machine learning part of computational statistics is usually thought of as supervised (e.g., classification) and unsupervised (e.g., cluster) methods. Beyond the mathematical and computational aspects of the machine learning methods, however, understanding what conclustions can be drawn from the methods is equally important. Cross validation helps the analyst work through the bias-variance trade-off. Feature engineering done well ensures that the information used in the model is the right information for the problem at hand.