Data science in insurance

Abstract

The globe is currently creating massive volumes of data, with data output in recent years increasing at an unprecedented rate. Most of this additional knowledge is being collected in novel ways, and technological developments allow it to be stored and analysed much more efficiently than usual. Against this background, there has been a lot of recent discussion about big data and data science. That is the capacity to analyse and derive valuable implications from increasing amounts of data from various sources quicker than ever. Data science is already transforming many facets of modern life, and it has enormous potential to foster innovations in the insurance sector. Insurers have traditionally collected information to understand premiums and risks better. Data science, coupled with increased computer capacity, provides a step-change in vulnerability assessment by allowing insurers to observe these risks in much greater depth continually. This can benefit insurers and policyholders, with room for innovations in how insurance is presented and priced and how claims are managed. As customers’ aspirations of all sorts of information, ability to respond quickly, and ways of conducting business rise, so will their expectations of the insurance sector. This chapter discusses many machine learning algorithms for effectively analysing insurance claims and comparing their performance using various criteria.