Statistical Machine Learning and Data Analytic Methods for Risk and Insurance

Gareth Peters

Research output: Working paper


Lecture series on Statistical Machine Learning for Risk and Insurance. PART I. Unsupervised Learning Methods In Risk and Insurance: The goal of this part of the lecture series is to develop core aspects of dimension reduction and feature extraction methodology from modern machine learning and data analytic approaches that can be directly applicable to working with Risk and Insurance modelling contexts. Feature Extraction and Dimension Reduction Methods in Risk and Insurance: In this section we will cover aspects of feature extraction methodology and dimension reduction methods that have the following core attributes: 1. capable of extracting relevant feature information that can be understood from a clear statistical perspective; 2. capable of treating outliers in the data (robust methods); 3. capable of treating missing data in the records; 4. computationally efficient to extract features. Detailed Real Applications to Illustrate the Ideas: a) Examples are developed in areas of: b) Premium and rate making applications in insurance c) Cyber risk and Operational Risk d) Claims reserving e) Home and contents insurance f) Mortality and Demographics in life insurance and annuities g) Telematics h) Catastrophe insurance i) Agricultural Insurance
Original languageEnglish
Number of pages309
Publication statusPublished - 2017


  • Insurance
  • Machine Learning
  • Risk
  • Statistical Learning


Dive into the research topics of 'Statistical Machine Learning and Data Analytic Methods for Risk and Insurance'. Together they form a unique fingerprint.

Cite this