Loss Models: From Data to Decisions, 4th Edition

4.1 Introduction

Basic probability models for actuarial situations tend to be either continuous or discrete (i.e., not mixed). The situation calls for either counting something (discrete) or paying something (continuous). In both cases it is unlikely the model will need to accommodate negative values. The set of all possible distribution functions is too large to comprehend. Therefore, when searching for a distribution function to use as a model for a random phenomenon, it can be helpful if the field can be narrowed.

In Chapter 3, distributions were distinguished by tail weight. In Section 4.2 distributions are classified based on the complexity of the model. Then in Chapter 5 a variety of continuous models is developed. Chapter 6 provides a similar treatment of discrete models.