Objectif du cours
This course introduces a range of modern research techniques to deal with high dimensional and potentially unstructured data. These techniques fall into two distinct categories. The first part of the course introduces statistical learning techniques. In recent years, there has been an increasing interest in predictive modelling, in particular for the collection of original research data that would not be available using traditional methods. This course provides a comprehensive understanding of some of the most capable supervised learning algorithms, including support vector machines, random forests, or neural networks. The second part of the course focuses on applied spatial economics. Many economic phenomena are spatial in nature. Manipulating and analysing spatial data rely on specific sets of tools known as geographical information systems. Specifically, students will learn how to manage vector and raster data, perform geocomputations, represent spatial processes and fit spatial models. Broadly speaking, the lectures use comprehensive theorising and mathematical formalisation but keeps a strong focus on intuition and effective implementation.
- R programming - Geographical information systems - Machine learning
Knowledge of :
- Econometrics: Advanced - R programming: Intermediate
Lecturer : Clément Gorin
This is a 8 week-course, typically with one 3 hours lecture each week.
1 - Statistical learning 2 - Advanced R 3 - Non-linear modeling 4 - Tree-based methods 5 - Support vector machines 6 - Neural networks 7 - Spatial data 8 - Geocomputations
- Individual report - 80%
- Critical review of a research paper - 20%
The report should be 10 to 12 pages long, and is due by December 6th, along with a script reproducing your results. Your work should articulate statistical learning techniques and/or geographical information systems, along with traditional econometrics. Grading is based on the following criteria:
- Question: Research question of interest to economists
- Background: Short review of the related literature
- Data: Collection and construction of the databases
- Model: Model selection, diagnostics and corrections
- Results: Meaningful interpretation and answer to the research question
- Writing: Clear and concise! English, referencing, etc.
You may use Rmarkdown to weave together text, code and output.
- R Athey, Susan. The impact of machine learning on economics. National Bureau of Economic Research, 2018.
- Bivand, Roger S, Edzer Pebesma, and Virgilio Gomez-Rubio. Applied spatial data analysis with R. Springer-Verlag, 2013.
- Gareth, James, Trevor Hastie, Robert Tibshirani, and Daniela Witten. An introduction to statistical learning with applications in R. Springer-Verlag, 2014.
- Goodfellow, Ian, Yoshua Bengio, and Aaron Courville. Deep learning. MIT Press, 2016.
- Hastie, Trevor, Robert Tibshirani, and Jerome Friedman. The elements of statistical learning. Springer, 2009.
- Sendhil, Mullainathan and Spiess Jann. Machine learning: An applied econometric approach. Journal of Economic Perspective, 31(2): 87-106, 2017.
- Venables, William N, David M Smith, and R Core Team. An introduction to R. Network Theory Ltd., 2009.
- Wickham Hadley. Advanced R. Chapman and Hall, 2019.