Course description

The objective of this course is to show students how statistics is used in practice to answer a specific question, by introducing a series of important model-based approaches.

The students will learn to select and use appropriate statistical methodologies and acquire solid and practical skills by working-out examples on real-world data sets from various areas including medicine, genomics, ecology, and others.

All analyses will be conducted mainly with the R software, possibly with interfacing to Python. No strong knowledge neither of R nor Python programming is required (only basic scripting).

Important remark

Much of the material used in this course is due to Marc Lavielle, who was the first to set up the Statistics in Actions course. We only have made some adjustments to it.

Schedule (tentative)

Teachers : Julien Chiquet (lecture + 1 PC), Geneviève Robin (2 PC)

Course Evaluation: 2 individual homework assignments + a final exam/project

Course Language: French with all material in English

  1. Statistical tests (x2)

    • Two-populations comparison
    • Power analysis
    • Multiple Testing
  2. Regression models (x2)

    • Linear and Non Linear Regression models
    • Nonlinear regression models
    • Inference Diagnostic, Model comparison
  3. Mixed effects models (x2)

    • Linear mixed effects models
    • Nonlinear mixed effects models
  4. Mixture models and model-based clustering (x3)

    • Gaussian mixture models for data clustering
    • Hierarchical and Spectral clustering for Graph
    • Stochastic Block Models
    • (Variational) EM algorithm