This training course enables the participants to reinforce their theoretical and practical knowledge in order to implement machine learning techniques for data analysis. The main algorithms for prediction (supervised learning) are presented. Each method is first presented and commented on a theoretical level, and then illustrated on numerical experiments run with public datasets using R and/or python/scikit-learn software.

To present the main algorithms for supervised learning : linear models for regression and classification, Classification and Regression Tress, Random Forests, introduction to Neural Networks and Deep Learning and to use them with R and/or python/scikit-learn.

This training session is for researchers, students, engineers who wish to reinforce or extend their theoretical background and practical knowledge on automatic data analysis by learning algorithms.

• Basic knowledge in statistics: elementary probability and statistics.
• Basic knowledge in algorithmic and programming.
• Install Python 3.6 with Anaconda.
• Install with conda the Keras library including TensorFLow.
• Install R and Rstudio and IR kernel
• Internet access during the sessions in order to get possible updates or to load additional libraries.

Scientific contacts: Béatrice Laurent-Bonneau, Olivier Roustant

• Day1 : Linear Models for Regression and Classification

  -9H-10H30: Lecture Linear Models for Regression Slides

  -10H30-11H : Coffee break

  -11H-12H30: Tutorials Ozone concentration prediction models

  Execute the first part of the two tutorials using respectively R and python/scikit-learn softwares.

  -14H-15H30 : Lecture Linear models for Classification Slides

  -15H30-16H : Coffee break

  -16H-17H30 : Tutorials Ozone concentration prediction models

  Execute the second part of the two tutorials dedicated to classification

• Day2 : Non linear Machine Learning algorithms

  -9H-10H30 : Lecture Classification and Regression Trees, Random Forests Slides

  -10H30-11H : Coffee break

  -11H-12H30 : Lecture Neural Networks and Introduction to deep learning Slides

  -14H-17H30 : Tutorials Human Activity Recognition

  Execute the Python tutorial where all the supervised classification algorithms studied in this training will be applied to Human Activity Recognition based on signals recordings obtained with a smartphone.

N.B. Notebooks analysing many other use cases are available on Wikistat.