Multivariate Exploratory Data Analysis Using Component Models

This chapter presents a comprehensive introduction to Exploratory Data Analysis (EDA) of omics data using component models. Many component models are introduced and explained starting from Principal components analysis (PCA), and its sparse variants (sparse PCA, and Group-wise PCA). The analysis of multilevel data is addressed through Multilevel Simultaneous Component Analysis (MLSCA), while designed data are explored with ANOVA-Simultaneous Component Analysis (ASCA) and its sparse variant Group-wise ASCA (GASCA). Analysis of covariance/correlation matrices is dealt with Covariance Simultaneous Component Analysis (COVSCA). Dimensionality assessment in PCA and data pre-processing are also discussed. The methods are applied and compared on experimental data.