Catherine Pardoux

Statistical Learning and Data Science

Data analysis is changing fast. Driven by a vast range of application domains and affordable tools, machine learning has become mainstream. Unsupervised data analysis, including cluster analysis, factor analysis, and low dimensionality mapping methods continually being updated, have reached new heights of achievement in the incredibly rich data world that we inhabit. Statistical Learning and Data Science is a work of reference in the rapidly evolving context of converging methodologies. It gathers contributions from some of the foundational thinkers in the different fields of data analysis to the major theoretical results in the domain. On the methodological front, the volume includes conformal prediction and frameworks for assessing confidence in outputs, together with attendant risk. It illustrates a wide range of applications, including semantics, credit risk, energy production, genomics, and ecology. The book also addresses issues of origin and evolutions in the unsupervised data analysis arena, and presents some approaches for time series, symbolic data, and functional data. Over the history of multidimensional data analysis, more and more complex data have become available for processing. Supervised machine learning, semi-supervised analysis approaches, and unsupervised data analysis, provide great capability for addressing the digital data deluge. Exploring the foundations and recent breakthroughs in the field, Statistical Learning and Data Science demonstrates how data analysis can improve personal and collective health and the well-being of our social, business, and physical environments.

Symbolic Approaches for Three-way Data

The increasing amount of information proposed to statisticians includes, in many applications, quantitative, qualitative or symbolic data which are observed at different time points t = 1,…, T. This implies three-way data, i.e. a sequence \({\underline X _1}\),…, \({\underline X _T}\) of T two-dimensional symbolic data arrays \({\underline X _t}\), = (X kit ) where k indexes individuals and j indexes variables. The investigation of such data requires to adapt and generalize classical and symbolic data analysis methods to the case of time series.