Description: This paper demonstrates the calculation of the moments of the distribution of aggregate life insurance claims from seriatim inforce data, assuming that each record represents an independent Bernoulli trial with its own known probability of a claim, q. Continuous distributions matching the first three or four moments are identified.

Description: Claims are insurers deepest and richest transactional data – yet we rarely unleash the full potential of this data. In this session, we will discuss potential sources and attributes of claims data, traditional approaches alongside new analytic techniques and how to integrate models into existing claims management processes. During this session, the presenters will discuss different areas of how machine learning can drive significant value unleashing the potential in claims data, these include data capture, claims management and workflow design as well as behaviour prediction. In addition, the audiences will also have opportunity to learn different use cases of how claims models taking advantage of machine learning, such as foreign death fraud detection and critical illness triage.

Description: This paper reviews recursive formulas for aggregate claim distributions. Followed by a submitted discussion paper

Description: The goal of this paper is to develop best possible estimates for the higher moments of a distribution of a positive bounded random variable, such as claim amounts, where these estimates are given in terms of the mean.

Description: The purpose of this paper is to introduce a practical method for interpolation and smoothing and give a number of numerical examples using the method. The method is a variation of the Whittaker-Henderson method. Discussions of this paper are included.