Outline

Day 1: Introduction, clustering methods and CART

Part 1: What can Predictive Analytics methods and Machine Learning be used for in Insurance?

1.1 Introduction
1.2 Software
1.3 Applications

Part 2: Clustering methods, classification and regression trees

2.1 Clustering and other unsupervised learning methods
2.2 Classification and Regression Trees (CART)
2.3 Applications: credit scoring, auto insurance, fraud detection, credibility

Day 2: The Random Forest and Stochastic Gradient Boosting algorithms

Part 3: Random Forest

3.1 The Random Forest algorithm
3.2 Interpretation of the results.

Part 4: Stochastic Gradient Boosting

4.1 The benefits of the Stochastic Gradient Boosting algorithms
4.2 How Stochastic Gradient Boosting algorithms work
4.3 Interpretation of Stochastic Gradient Boosting output.

Day 3: Multivariate Adaptive Regression Splines

Part 5: Multivariate Adaptive Regression Splines (MARS)

5.1 Introduction to non-linear regression with splines
5.2 The MARS algorithm
5.3 How to use MARS models.

Part 6: Application to credit scoring and insurance problem

Day 4: Risk management for variable annuities

Part 7: Motivational example: variable annuities

7.1 Cash flows
7.2 Risks associated with the guarantees

Part 8: Models for risks underlying variable annuities

8.1 Equity risk models
8.2 Interest rate models
8.3 Mortality risk
8.4 Policyholder behavior risk
8.5 Basis risk

Day 5: Pricing and hedging for variable annuities II

Part 9: Pricing and reserving methods for variable annuities

Part 10: Hedging methods for variable annuities

10.1 Hedging instruments
10.2 Review of various hedging approaches

Part 11: Reinforcement Learning in Finance

11.1 Introduction to Reinforcement Learning
11.2 Applications to variable annuities