PhD course: An Introduction to Functional Data Analysis
Theory and Practice
Date: Monday March 25th to Thursday March 28th 2024
Place: dSEAS, University of Palermo, Italy
Welcome!
Course Description
Functional Data Analysis (FDA) is a branch of statistics focused on the analysis and interpretation of data on a continuum, meaning that they can be described naturally as smooth curves or functions. Recent technological advances allow collection of data at such high-frequency that classical multivariate statistical techniques are not able to fully capture and describe their complexity, resulting in the flourishing of functional data methodology. FDA can easily be thought of as the collection of all statistical methodologies needed to handle samples of curves. These include mostly classical multivariate statistical methods adapted to the infinite dimensional case, such as regression models, inference, dimensional reduction, and clustering. However, methods that are specifically needed in this context have also been developed, such as data smoothing and registration. FDA methods have been applied quite broadly, from medicine to engineering, from business, to demography and social sciences, affecting nearly all fields of science. This course offers an introduction to FDA from a methodological viewpoint and gives practical guidance to R packages used in such context.
Course Objective
The overall objective of the course is that the PhD students in statistics at the DSEAS of the University of Palermo (Italy) develop skills in FDA, both at the theoretical (understanding the methodologies) and practical level (actual implementation in R). At the end of the course the students should be able to:
- identify situations in which they can treat their data as functional
- apply standard FDA techniques (data description, dimensionality reduction, clustering, regression, and inference)
- visualize the results
The course will last 4 days, comprising lectures connected to practical hands-on sessions.
Exam
Evaluation of the course learning outcome will be carried out via a take-home exam (see the dedicated page for details)