Duration
30h Th
Number of credits
Lecturer
Language(s) of instruction
English language
Organisation and examination
Teaching in the first semester, review in January
Schedule
Units courses prerequisite and corequisite
Prerequisite or corequisite units are presented within each program
Learning unit contents
This course is a statistical (data) modelling course, aiming at providing students with a fundamental knowledge of regression and statistical modelling techniques with applications in the fields of Finance and Economics, as well as being able to solve these problems with the help of the software R. It ambitions to give students a firm basis in data analysis and statistics for their master thesis. It is also of interest for students aiming at a PhD in economics/finance or wanting to prepare the FRM certification.
The course will mix statistics, financial econometrics and financial engineering classes. Starting from practical data modelling problems, I will explain the types of statistical models that can be used to solve the problems and the underlying statistical theory. Then, during practical sessions, the students are expected to work by themselves on related exercises, with the aim to eventually conduct realistic data analyses. During these practical session, a teaching assistant will be available for questions.
Regarding the statistics/financial econometrics part, I will cover topics related to classical and advanced statistics/regression techniques (linear regression, Generalized Linear Model) and machine learning methods (LASSO, neural network, random forest). More philosophical notions related to quantitative studies in social sciences, and details on how to conduct prediction/classification tasks will be covered through selective readings.
Regarding applications, the course focuses on topics related to predictive regression analysis (for example linear and nonlinear models) such as statistical arbitrage techniques, alpha or return predictions, classification problems such as default prediction, and optimal portfolio allocation with machine learning methods. This list is open to suggestions and will evolve according to students' proficiency and interests.
The final grade consists mainly in a data analysis project per group of (around) two students. The project may be choosen by the students (after acceptance by me) and motivated by an existing research question in the financial or economic literature, or imposed by the lecturer. In the former, examples include the replication of an existing paper or the first step of a master thesis. In the latter case, examples of previous projects were related to realized volatility prediction or mutual fund selection with machine learning. The students will have to present their project in front of the class during a seminar session taking place at the end of the semester. There is no written exam in January. This grade may be complemented by the presentation of a scientific article and its discussion during the semester (details will be announced at the beginning of the class). The article may cover advanced asset pricing and statistical arbitrage topics in relation with textual analysis, autoencoders, multiple testing. Finally, a bonus will be granted based on the performance during a trading competition with the platform StockTrak organized throughout the semester.
For the second session, students are expected to conduct an imposed replication exercise of a research article, and to present it during an oral exam. At that occasion, theoretical knowledge of the course will be stressed. Grades of the assignment and the trading competition are not transferable to the second session.
The course is taught in English.
Learning outcomes of the learning unit
At the end of this course, the students are expected to:
- have a theoretical understanding of advanced regression and statistical learning methods, and to be able to explain the theoretical foundations of these models,
- understand the notions of model selection and high dimensionality problems, and to be able to explain these concepts,
- identify the various situations and types of data for which these models are useful (binary or categorical outcomes, daily time series, ratio, etc.), and be able to explain why they chose a particular model to analyse these data.
- be able to write scripts in R, using existing routines or implementing their own solution, to estimate these models and conduct a data modelling exercise,
- be able to read and understand scientific methodological articles, to implement the proposed methodology.
In addition, the students will develop the following competences:
- use of a foreign language (English, with emphasis on a scientific vocabulary),
- writing of a term paper,
- ability to present oraly scientific concepts,
- ability to work autonomously.
Prerequisite knowledge and skills
- Advanced knowledge and interest for statistics and programming. In particular, a good understanding of bachelor courses like "Probabilite et Inference statistiques", "Models and Methods in Applied Statistics" or "Econometrics" is a prerequisite for the statistical side.
- Courses or self-teaching of the fundamentals of programs like R, VBA or Python is a good basis. Learning material for a crash-course in R will be provided, relying on the learning platform Datacamp (https://www.datacamp.com/)
- Courses like Advanced Econometrics, Empirical Methods in Financial Markets, Seminar in Applied Econometrics, Financial Risk Modelling are good pre-requisites or complements.
Planned learning activities and teaching methods
Ex-cathedra lectures (theory), exercises, oral presentation during seminars, self reading of scientific articles.
Mode of delivery (face to face, distance learning, hybrid learning)
Face-to-face course
Further information:
Face-to-face.
Course materials and recommended or required readings
Platform(s) used for course materials:
- LOL@
Further information:
The main references will be:
The Elements of. Statistical Learning: Data Mining, Inference, and Prediction. Second Edition. February 2009. Trevor Hastie Robert Tibshirani.
An Introduction to Statistical Learning: with Applications in R, by G. James et al. (2013).
Regression: Models, Methods and Applications. Berlin: Springer-Verlag. by Fahrmeir, Ludwig; Kneib, Thomas; Lang, Stefan; Marx, Brian (2013).
Empirical asset pricing via machine learning S Gu, B Kelly, D Xiu The Review of Financial Studies 33 (5), 2223-2273
Financial machine learning, B Kelly, D Xiu, Foundations and Trends in Finance 13 (3-4), 205-363
There will be complemented by slides.
Exam(s) in session
Any session
- In-person
oral exam
Written work / report
Out-of-session test(s)
Further information:
- Participation to the intermediate flipped classroom session, where students are expected to present a research article and lead a discussion on this topic.
- Oral presentation (around 20 min + questions, depending upon the number of participants) of the data analysis project during a final seminar.
Work placement(s)
Organisational remarks and main changes to the course
Contacts
Prof. Julien Hambuckers
email: jhambuckers[at]uliege.be
Teaching assistant: P. Hübner (phubner[at]uliege.be)