Machine learning in finance

FIN-407 / 6 crédits

Enseignant: Malamud Semyon

Langue: Anglais

Summary

This course aims to give an introduction to the application of machine learning to finance, focusing on the problems of portfolio optimization, return prediction, and textual analysis. A particular focus will be on deep learning and the practical details of applying deep learning models to finance.

Content

 

1- Introduction to machine learning in finance

  • Goals of machine learning
  • Applications of machine learning
  • Optimizaing algorithms and Inductive Biases

2. Regression,

  • ridge regression
  • overfitting
  • sparsity
  • penalization

2- Neural Networks

  • Training, testing, overfitting, and penalizing
  • Feedforward networks
  • Gradient descent dynamics, catapults, fine-tuning

3- Introduction to Natural Language Processing

  • Text representation and Embeddings
  • Sentiment analysis
  • Large Language Models (LLMs)

4. Transformers

  • Transformer architecture
  • Attention
  • Masking
  • Applications to LLMs

5. Portfolio Optimization

  • Risk-Return tradeoffs, conditional and unconditional efficiency
  • Deep and Shallow Learning, Factors, ML portfolios
  • Transformers for portfolio optimization and construction

 

 

 

 

Keywords

Machine Learning, Deep Learning, NLP, Python

Learning Prerequisites

Required courses

Introduction to Econometrics

Basic knowledge of Python. If you do not know Python, please take

https://www.my-mooc.com/fr/mooc/programmation-en-python-pour-debutants

or any other introductory Python class.

Warning: 90% of the class is based on coding in Python

Recommended courses

Introduction to finance

Important concepts to start the course

Basic linear algebra.

Basic probabilistic and statistical concepts.

Basic knowledge of Python.

 

Learning Outcomes

By the end of the course, the student must be able to:

  • Elaborate a machine learning algorithm
  • Assess / Evaluate the performance of different models
  • Formulate hypotheses behind different models
  • Propose optimal methods for problems seen
  • Optimize techniques / algorithms used
  • Construct a parsimonious model
  • Implement machine learning algorithms
  • Exploit information contained in data

Transversal skills

  • Give feedback (critique) in an appropriate fashion.
  • Demonstrate the capacity for critical thinking
  • Use a work methodology appropriate to the task.

Teaching methods

Lectures and exercise sessions

Jupyter Notebooks

Projects (coding in Python)

Expected student activities

  • Participate in lectures
  • Participate in exercises sessions (coding in Python)
  • Solve the problem sets (coding in Python)
  • Work on a project (coding in Python)
  • Final exam: coding in Python

Assessment methods

0.1 * Class Participation + 0.3 *Project report + 0.2 * homework + 0.4 * Exam

Supervision

Assistants Yes

Resources

Bibliography

Dixon M. F, Halperin I. and Bilokon P. (2020): "Machine Learning in Finance", Springer

De Prado, M. L. (2018). Advances in financial machine learning. John Wiley & Sons.

Basic knowledge of Python. If you do not know Python, please take

Ressources en bibliothèque

Moodle Link

Prerequisite for

  • Courses using statistical dynamic models

Dans les plans d'études

  • Semestre: Printemps
  • Forme de l'examen: Ecrit (session d'été)
  • Matière examinée: Machine learning in finance
  • Cours: 3 Heure(s) hebdo x 14 semaines
  • Exercices: 2 Heure(s) hebdo x 14 semaines
  • Type: obligatoire
  • Semestre: Printemps
  • Forme de l'examen: Ecrit (session d'été)
  • Matière examinée: Machine learning in finance
  • Cours: 3 Heure(s) hebdo x 14 semaines
  • Exercices: 2 Heure(s) hebdo x 14 semaines
  • Type: obligatoire
  • Semestre: Printemps
  • Forme de l'examen: Ecrit (session d'été)
  • Matière examinée: Machine learning in finance
  • Cours: 3 Heure(s) hebdo x 14 semaines
  • Exercices: 2 Heure(s) hebdo x 14 semaines
  • Type: optionnel

Semaine de référence

Cours connexes

Résultats de graphsearch.epfl.ch.