Block Course „Deep Learning with Actuarial Applications in R”
Purpose and Nature:
The goal of this two-day block course is to introduce the participants to neural networks for the modelling of insurance risk in general and life insurance. Therefore, we will provide an introduction into neural networks from a mathematical perspective and from an implementation point of view using R and the keras package. We also give a short introduction into unsupervised learning.
The course will be practical and consists of approx. 50% presentations where the models are introduced and explained. The other 50% are exercises where the participants are using R/RStudio, the keras package and apply the concepts to publicly available insurance risk data.
The content of the course is mainly based on the published tutorials of the SAV “Data Science” working group and available on https://www.actuarialscience.org.