Analysis of three-way data: an extension of the STATIS method

In order to explore differences and similarities in three-way data (units × variables × occasions), an extension of the STATIS method is presented. STATIS is a generalization of PCA to analyze several sets of variables collected on the same units. As a novelty, the method proposed here searches for a “compromise” of the occasions by assuming that the variables can be differently weighted to capture the similarity structure between the corresponding variables across the occasions. An application to real data is presented to illustrate the potentiality of the method.