Data science - how actuaries can optimise usage of data science techniques within their sectors 2020
Machine learning is widely used in general insurance pricing, where it has been shown that techniques such as GBMs and neural networks outperform the previous benchmark of GLMs. Since it is well known that reserving may be expressed as a GLM problem, this then suggests that machine learning could also be used to deliver superior insights and/or results. Although not yet widely used in reserving, there have been a number of interesting papers over recent years, some of which are discussed here.