21 January 2020
Actuaries are increasingly looking to explore data science techniques as a way to deliver new insights, utilise new datasets and develop complex models efficiently. However challenges remain to integrate these new techniques into the standard actuarial toolkit. A key challenge for actuaries is to understand the steps involved in a typical data science project, including how to create a robust framework for developing and reviewing advanced statistical models. This paper presents an overview of the steps in a typical data science process and a worked case study provides a practical example the approaches taken to validate a machine learning model.