Tutorial: Multiway Data Analysis
Sébastien Lê, Agrocampus Rennes, Rennes Centre for Higher Education and Research in Agronomy, Rennes Cedex, France.
Julie Josse, Agrocampus Rennes, Rennes Centre for Higher Education and Research in Agronomy, Rennes Cedex, France.
François Husson, Agrocampus Rennes, Rennes Centre for Higher Education and Research in Agronomy, Rennes Cedex, France.

Motivation

In many frameworks, the researcher is interested in the relationship between several sets of variables. The aim of this tutorial is to propose an overview of exploratory methods that handle multiway data (Generalized Canonical Analysis, Multiple co-inertia analysis, Multiple Factor Analysis, Generalized Procrustes Analysis, etc) and to stress on a particular method, Multiple Factor Analysis.
After the tutorial the participants will be able to be autonomous to explore multiway data sets.

Outline

  1. Introduction to multiway data tables via genomics, sensory and ecological data and their specific problematics
  2. Overview of statistical methods which take into account a partition on the variables
  3. Some features of Multiple Factor Analysis
    1. The interest to balance the influence of each group of variables
    2. Representation of the information (by group of variables, by individuals, by partial points, etc) in a unique framework
    3. How to add supplementary information? Supplementary information does not participate to the construction of the analysis but is a tool to facilitate the interpretation of the analysis
    4. Classical example with sensory data (wines described by several types of information: visual aspects, olfoctary feelings, gustative appreciations, ...)
  4. Advanced uses of Multiple Factor Analysis
    1. On the interest of using MFA before an unsupervised classification
    2. Comparison of coding (continuous or categorical) with the use of MFA
    3. Integrating `omics' data sets and biological knowledge via MFA

Background knowledge

A basic statistical knowledge is required on principal component analysis

Intended audience

Teachers in data mining and data analysis, researchers in applied field, statisticians whose topic of interest is multivariate analysis

Workshop materials

More information will be available (notes, scripts, and data sets) at our website (http://factominer.free.fr)