Economic and Financial Applications of Machine Learning
The module focuses on developing, deploying and testing machine learning models for real-world financial decisions. Applications include credit risk, insurance fraud, customer relationship management, factor investing and portfolio construction, with attention to data preparation, model tuning and robust backtesting.
Prerequisites
To take this module you must first have completed Advanced Machine Learning and Programming in Python.
Delivery
Live online lectures and tutorials, conducted at scheduled times and recorded for review.
Study Hours
Approximately 100 hours of study, of which 15 are contact hours and 85 are private study.
Software Used
Python
Credits
10 CATS credits
Module Lead
Programme Faculty
Key Outcomes
- —Apply machine learning to economic and financial problems
- —Develop, deploy and test predictive models
- —Work with credit risk, fraud, churn and portfolio applications
- —Prepare real-world data and tune model parameters
- —Design robust backtesting procedures
Desired Skills
- —Analyse and interpret real economic and financial data
- —Solve complex applied problems using machine learning
- —Explain model choices and results clearly
Lecture Programme
- WK 01
Credit risk management
- ·Developing a credit scorecard
- ·Basel II and III regulation
- ·Probability of default modelling
- WK 02
Insurance fraud detection
- ·Predictive analytics
- ·Social network analytics
- WK 03
Customer churn and response modelling
- ·Causal machine learning
- ·Profit-driven modelling for CRM
- WK 04
Foundations of factor investing
- ·Fama–French models
- ·Portfolio sorts
- WK 05
Data preparation and penalised models
- ·Missing data
- ·Lasso
- ·Sparse portfolios
- WK 06
Nonlinear models
- ·Tree-based methods
- ·Neural networks
- WK 07
Parameter tuning and backtesting
- ·Grid search and overfitting
- ·Machine-learning-powered backtests
Assessment
Online coursework submission at the end of the module.
Build your qualification