This lecture will introduce popular unsupervised approaches to modelling high dimensional data i.e. data for which a large number of features are available for each observation. The mathematical underpinnings of principal components analysis (PCA) will be introduced, as will PCA’s links to probabilistic latent variable models. The lecture will be complemented by a hands on session, allowing attendees to apply and use the introduced methods using the R statistical computing environment.
This lecture will include breakout sessions and particpants should have downloaded the following software: R statistical computing environment: www.r-project.org and RStudio: rstudio.com
4-5pm: Dr Norma Bargary, UL, From raw data to smooth curves: an introduction to functional data analysis
Brief synopsis of lecture content:
Vast amounts of data are now collected at an enormous rate. For example, data from sensors are now routinely measured across diverse fields including medicine, climatology, ecology, environmental science, sports science, manufacturing and the Internet-of-Things (IoT). The analysis of these large, complex datasets is a key topic of modern statistics from both a methodological and practical perspective. Functional data analysis (FDA) is a statistical framework that treats each measured data stream as a realisation of a smooth, differentiable function, thus exploiting key properties of the data. A key first step in FDA involves smoothing to remove noise and produce a finite-dimensional representation of the data for model fitting, which is then subsequently analysed. This lecture will introduce smoothing methods and other FDA methods that can be used to analyse high-dimensional, high-throughput data.
This lecture will include breakout sessions. Participants need laptops with R and RStudio installed plus the fda R package installed. Alternatively participants can work on RStudio cloud (create an account for free on https://rstudio.cloud ) and all material will be available there. This option does not require any R/RStudio to be installed on laptops locally.
Claire is a Professor in the UCD School of Mathematics and Statistics where her research focuses on developing apposite statistical methods for the analysis of complex data arising in real world settings. Her research develops novel statistical methods, largely based on latent variable models, for the analysis of high dimensional data, often of mixed type. The methods solve applied problems across a range of disciplines, including metabolomics, genomics, social science, food science, sports science and political science.
Dr Norma Bargary is a lecturer in Statistics at the Dept. of Mathematics & Statistics, UL. She is a member of MACSI, Ireland’s foremost industrial and applied mathematics/statistics research group. Her primary research interests are the statistical modelling of high-dimensional data, and data measured as curves (functional data analysis). She has extensive experience working with collaborators across diverse disciplines (medicine, cancer research, sports science, psychology, biomechanics) and industry partners (manufacturing, pharma). She has been the statistical lead in several successful European and National grant applications and is vice-director of CRT. She has a keen interest in public engagement and is the first of two Royal Statistical Society Statistical Ambassadors based in Ireland.
The SFI Centre for Research Training in Foundations of Data Science will train a cohort of PhD students with world-class foundational understanding in the horizontal themes of Applied Mathematics, Statistics, and Machine Learning.
Let us be part of your #PhD journey. 36 fully-funded studentships waiting to be filled starting in Sept 2021. Hear what our students say about the CRT @UCDGS @UL @MaynoothUni @IUAofficial #believeinscience #phdchat
For perfomance reasons we use Cloudflare as a CDN network. This saves a cookie "__cfduid" to apply security settings on a per-client basis. This cookie is strictly necessary for Cloudflare's security features and cannot be turned off.