401-3936-00L  Data Analytics for Non-Life Insurance Pricing

SemesterSpring Semester 2018
LecturersC. M. Buser, M. V. Wüthrich
Periodicityyearly recurring course
Language of instructionEnglish


AbstractWe study statistical methods in supervised learning for non-life insurance pricing such as generalized linear models, generalized additive models, Bayesian models, neural networks, classification and regression trees, random forests, gradient boosting machines and support vector machines. Moreover, we present unsupervised learning methods applied to telematics car driving data.
ObjectiveThe student is familiar with classical actuarial pricing methods as well as with modern machine learning methods for insurance pricing and prediction.
ContentWe present the following chapters:
- generalized linear models (GLMs)
- generalized additive models (GAMs)
- credibility theory
- classification and regression trees (CARTs)
- bagging, random forests and boosting
- support vector machines (SVMs)
- unsupervised learning methods
- telematics car driving data
Lecture notesThe lecture notes are available from:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2870308
Prerequisites / NoticeThis course will be held in English and counts towards the diploma of "Aktuar SAV".
For the latter, see details under www.actuaries.ch

Good knowledge in probability theory, stochastic processes and statistics is assumed.