In recent times there is a growing need for models to handle big data problems. In machine learning techniques, there is an increasing use of GLM as a unified estimation and prediction method. There are different ways to use GLM in big data analytics. One approach is to use a regularized GLM to control for over or underfitting of the model. The traditional regression or classification techniques may result in large variance that can be controlled with regularization at the cost of larger bias. Decision trees and classification and regression trees (CART) can be combined with GLM as linear combinations to develop a more stable strategy for predictions with higher predictive accuracy. The flexibility of GLM in dealing with both discrete and continuous outcome variables under a unified framework makes it possible to employ GLM in determining optimal splits. Chapter 11 provides the necessary background to combine the machine learning techniques with GLM to improve the prediction performance.