Machine Learning : empirical applications for finance (prerequisite : Python for Finance) (This course corresponds to the bloc 2/3 of the Certificate "Fundamentals of Data Science")
Ects : 3
Enseignant responsable :
- HOUCINE SENOUSSI
Description du contenu de l'enseignement :
Basics of ML
- Definitions, approaches and applications.
- Data mining (DM) : definitions and links with ML.
- Classification and regression problems.
- Building and evaluating an ML model.
- Presentation of the main approaches of ML/DM.
- Application I.
Decision Trees :
- Definitions and algorithms.
- Advanced methods based on DL : Bagging, Boostring and Random forests.
- Application II : Making a decision in finance.
Neural networks:
- Definitions.
- Learning in NN : grandient descent and Backpropagation.
- Advanced methods based on NN (Deep learning).
- Application III : : Stock pricing.
Reinforcement Learning :
- Definitions : Agents and environnments.
- Markovian Decision Process (MDP).
- Policies and optimal policies.
- Q-learning.
- Application IV : Trading.
Pré-requis obligatoires :
Python programming language.
Coefficient : 1Compétence à acquérir :
Building Machine Learning (ML) models for Finance problems. Using ML Python library (and in particular sickit-learn).
Mode de contrôle des connaissances :
Two/Three assignments (building a model + Python programming).