In recent years, predictive modelling has changed important aspects of actuarial practice.  Predictive modelling enhances traditional actuarial models with modern statistical tools and analysis. It uses emerging volumes of data to provide important insights into life and health insurance business, including how to identify appropriate risks, manage the risks insurance companies face, and improve the accuracy of actuarial projection models.

Predictive modelling impacts many areas of life and health insurance – from underwriting, to risk identification, to assumption setting and financial modelling. In underwriting, predictive modelling can be used to select the policyholders that meet desired risk profiles and to improve the accuracy of risk classification schemes.  For existing blocks of business, predictive modelling allows actuaries to identify key risk factors impacting financial performance, creating opportunities to proactively manage those risk. As a projection tool, actuaries use predictive modelling to identify the key factors impacting actuarial assumptions, and to appropriately fit the assumptions used in financial projection models to historical data, potentially improving the accuracy of actuarial projections.

As actuaries enhance their focus on and knowledge of predictive analytic tools, it is likely that predictive modelling will play an increasingly important role in actuarial practice in the future. This web session is designed to provide actuaries with the technical tools needed to be prepared for the ways in which actuarial practice may evolve in the coming years.

Organised by the EAA - European Actuarial Academy GmbH.

The web session is intended as a technical introduction for actuaries wishing to develop fundamental skills in predictive modelling techniques. It is designed for those who will be directly involved in the construction and analysis of predictive models. The web session presumes no prior knowledge of predictive modelling, but attendees should have a working knowledge of basic statistics, linear regression, and life and health actuarial models.

Attendees are encouraged to bring a laptop computer with Microsoft Excel and R installed.  The web session will presume a working knowledge of statistics and linear regression. R will be used only to a limited extent to demonstrate certain concepts and techniques; lack of familiarity with R should not prevent one from attending the session.  

Technical requirements
Please check with your IT department if your firewall and computer settings support web session participation (the programme GotoTraining/GotoWebinar is used for the web session). Please also make sure that you are joining the web session with a stable internet connection.

The web session will cover a broad range of topics in predictive modeling that are relevant to actuarial practice. The session begins with a review of multivariable linear regression, and relatively quickly transitions to more advanced techniques. Advanced topics include logistic regression, gradient boosting models (GBM), and generalised linear models (GLM).

Case studies will be used throughout the session to demonstrate the concepts in a practical setting. The case studies will focus on applications to life and health insurance.

Andrew H. Dalton
Andrew is a Principal and Consulting Actuary in Milliman’s Philadelphia office. Andrew’s professional experience includes work on actuarial appraisals for mergers and acquisitions, asset and liability analysis, predictive modeling, and economic capital for life and health companies. Andrew is a Fellow of the Society of Actuaries and a Member of the American Academy of Actuaries. He holds a Masters Degree in Business Administration, concentrating in Finance and Statistics, from the Leonard N. Stern School of Business of New York University.

Remi Bellina
Remi is an actuary of the French Institute of Actuaries and holds an engineering degree in Applied Mathematics. Remi has been a consultant at Milliman in Paris since 2012 after having worked one year at AXA Liabilities Managers. Since he joined Milliman, Remi has worked on many P&C topics including reserving, Solvency II ratio calculation (both Standard Formula and Internal Model), IFRS 17 projections and M&A projects. At the same time, Remi has carried out a number of pricing missions, by modeling the tariff or policyholder value on motor and home insurance products. Among other projects, Remi created motor tariffs on the liability cover for a fleet by using GLM approaches, he modernized existing tariffs by using innovative machine learning approaches (such as gradient boosting), and he gave several training sessions. As the Chief Data Scientist at Milliman, Remi has also a record of experience in using machine learning techniques, he wrote a master's thesis and other papers on using advanced algorithms and visualizations.

Floriane Moy
Floriane is a senior data scientist and has more than 5 years of experience in data analytics, statistical modeling and supervised and unsupervised machine learning. She holds an engineering degree in Applied Mathematics and Statistics and a master in Financial Engineering from Berkeley University. Floriane joined the Analytics Department of Milliman in 2019 as a Senior Consultant and Data Scientist. She led several projects for French insurers such as the implementation of a motor pricing, the modeling of client value, the design of scoring models to increase multi-equipment rates, the review of XG-Boost models for credit scoring, etc. She is highly proficient in implementing complex machine learning algorithms, in using open source languages for data processing and quantitative analysis such as Python or R and in developing tools (web applications, APIs, etc.). Prior to joining Milliman, Floriane worked in Asset Management in the US where she led research projects on portfolio diversification and prediction modeling for stocks performance.