DESCRIPTIVE MULTIDIMENSIONAL STATISTICAL-METHODS FOR ANALYZING SIGNALS IN A MULTIFACTORIAL BIOMEDICAL DATABASE

A methodology is presented to analyse multidimensional signals from se veral recording periods resulting from an experimental study on human or other living systems. The methodology is divided into two stages: i ntra-period analysis and inter-period analysis. The purpose of the fir st stage is to highlight general trends in multidimensional signal cha nges and the more informative components of the signals. The purpose o f the second stage is to assess the influence of environmental or indi vidual difference factors on a given signal component that appears to be discriminant in the first stage. To take into account the multivari able state of the system and the multi-observational aspect, a multidi mensional descriptive statistical approach is used. The methods are co rrespondence analysis and hierarchical clustering. They are illustrate d through an occupational medicine application from a study of sedenta ry posture.