Description: Reflection on three years on the PAF section council and where we're headed next.

Description: An introduction of predictive modeling techniques that are applicable to insurance industry, including generalized linear model (GLM) and classification and regression tree model(CART). Other models such as data clustering, neural network, and support vector machine (SVM), are also briefly discussed.

Description: A discussion on using the Cox Proportional Hazards Model in developing a predictive underwriting model that produces a mortality multiplier for each individual.

Description: A discussion about how many models can address the over fitting issues caused by too many variables by regularization, variable selection is still meaningful.

Description: This article provides an overview of activities that should be considered by an actuary when validating a predictive model. It describes several components to consider when validating a predictive model and provides an examination of various methodologies that can be implemented in the validation process.