ABSTRACT. In this paper we explore the relationship between multivariate data analysis and techniques for graph drawing or graph layout. Although both classes of techniques were created for quite different purposes, we find many common principles and implementations. We start with a discussion of the data analysis techniques, in particular multiple correspondence analysis, multidimensional scal...

per optar al grau de Doctora en Informàtica per Amalia Duch Brown sota la direcció del doctor Conrado Martínez Parra Barcelona, 25 d'octubre de 2004 Aquesta tesi fou llegida el dia 9 de desembre de 2004, davant el tribunal de tesi format per: Let no one say that I have said nothing original, at least the arrangement of the subject is new. RESUM iii Aquesta tesi està dedicada al disseny i a l'an...

In the following, I shall try to summarise developments over the past 100 years that have led to the current interest in data analysis and more particularly in multidimensional data analysis (see the Conclusion section for a discussion of these terms). Necessarily, my views are coloured by my own experience but I hope that some points of general interest will emerge. Developments may be divided...

Statistical units described by interval-valued variables represent a special case of Symbolic Objects, where all descriptors are quantitative variables. In this context, the paper presents two different metrics in R for interval-valued data that are based on the definition of the Hausdorff distance in R. Hausdorff distance in R (for any p ≥ 1) is a L∞ norm between pairs of closed sets. However,...

Cluster analysis techniques are used to classify objects into groups based on their similarities. There is a wide choice of methods with diierent requirements in computer resources. We present the result of a fairly exhaustive study to evaluate three commonly used clustering algorithms, namely, single linkage, complete linkage, and centroid. The cluster analysis study is conducted in the 2 dime...

Relatively recent strides in noninvasive medical imaging and in vital microscopy have yielded a new form of massive data type: the multi-channel, 3-D, time-course image recording. These multidimensional datasets document dynamic changes within the full volume of a specimen over time, often simultaneously monitoring several different parameters. As a result, visual data collected from a single l...

Understanding the functional implications of changes in gene expression, mutations, etc., is the aim of most genomic experiments. To achieve this, several functional profiling methods have been proposed. Such methods study the behaviour of different gene modules (e.g. gene ontology terms) in response to one particular variable (e.g. differential gene expression). In spite to the wealth of infor...