Sara Fontanella

I joined the Department of Medicine in November 2016 as a Research Associate in Statistical Machine Learning.

My research interests are Multivariate Statistics and Data Analysis. In particular, the main themes concern dimensionality reduction methods for high-dimensional data and sparse latent variable models.

I did my Ph.D. at the School of Advanced Studies G. d'Annunzio, Chieti (Italy). My thesis dealt with “Graphical models for spectral nonlinear dimensionality reduction”, and the main aim was to infer the possible relationships existing in high-dimensional data.

Prior to joining Imperial College, I was a postdoc at The Open University. I was working on a three year project entitled “Sparse factor analysis (FA) with application to large data sets” under the supervision of Dr N. Trendafilov. The aim of the research was to develop new approaches for Sparse FA. Several sparsity-inducing approaches. We applied both L1 penalty terms and approaches based on Bayesian statistics and Markov chain Monte Carlo methods, which are very useful especially in the context of high-dimensional data.

Currently, I am working within the multidisciplinary Study Team for Early Life Asthma Research (STELAR). My role is to implement and develop innovative computational statistical methods to identify novel subtypes of childhood asthma, enabling investigation of subtype-specific environmental and genetic associates and discovery of distinct pathophysiological mechanism.

I am also interested in the development and application of exploratory multivariate statistics and machine learning techniques in the context of zero-inflated data.