Tomas Rubin

About

My path from a mathematics student, through working in a tech start-up and business consulting, to becoming a researcher in statistics and data scientist

Biography

I'm a researcher in mathematical statistics working on developing novel methods for time series data analysis.

I'm finishing my PhD degree in statistics at the Swiss Federal Institute of Technology in Lausanne, Switzerland, (Ecole polytechnique fédérale de Lausanne, EPFL), a leading Swiss technical university. My PhD thesis introduces a novel statistical method to analyse time series data and spatial-temporal data. The novelty lies in removal of some assumptions of the existing methods, adaptability to a large number of missing values, and reduction of the prediction error up to 60 %. During the development of the method I was responsible for the design of the methodology, its implementation in Matlab, and its theoretical and simulation based justifications. I've successfully applied my methodology for meteorological data and in yield curve modelling.

Before joining the PhD programme at EPFL I gained experience in business consulting industry as well as in the Czech tech start-up scene.

Learn more about my endeavours and extracurricular interests by clicking on "Read more".

My projects, publications, and software

During the years at EPFL I was involved in my core research revolving around sparsely observed functional time series as well as in a few side projects.

A quantitative comparison of yield curve models in the MINT economies.

Posted in Publications by Tomas Rubin on August 20, 2020

A semester project I supervised at EPFL comparing various econometric forecasting methods for yield curves in MINT (Mexico, Indonesia, Nigeria, Turkey) economies.

Spectral Simulation of Functional Time Series

Posted in Publications by Tomas Rubin on July 18, 2020

A novel simulation method allowing for generating a wide range of simulated data, accompanied by an R package ‘specsimfts’.

Yield curve and macroeconomy interaction: evidence from the non-parametric functional lagged regression approach

Posted in Publications by Tomas Rubin on July 7, 2020

In my first single-author publication I applied the tools I developed in my core PhD research to yield curve modelling.

Sparsely Observed Functional Time Series: Estimation and Prediction

Posted in Publications by Tomas Rubin on July 3, 2020

The core paper of my PhD research. Published in Electronic Journal of Statistics (2020)

Functional Lagged Regression with Sparse Noisy Observations

Posted in Publications by Tomas Rubin on June 3, 2020

The paper extending results from my core paper on sparsely observed functional time series. To appear in Journal of Time Series Analysis

Stochastic Evolution Systems and Their Applications

Posted in Publications by Tomas Rubin on June 3, 2016

My Master’s thesis in stochastic analysis (probability theory) defended at Charles University in Prague, Czechia (2016)

View All

My recent activity

Events and stories in my professional path

Defended my PhD thesis!

Posted in Activity by Tomas Rubin on September 23, 2020

I successfully defended my PhD thesis on September 17, 2020, at Swiss Federal Institute of Technology in Lausanne (EPFL). The president of my jury, Sofia Olhede, and the examiners…

European Meeting of Statisticians (EMS 2019)

Posted in Activity by Tomas Rubin on July 30, 2019

In July 2019 I had the opportunity to travel to Palermo, Italy, to give a talk at the European Meeting of Statisticians, the largest statistics conference in Europe. In my talk entitled “Sparsely observed functional time series: estimation and prediction” I spoke about my first paper carrying the same name.

Doctoral day, Fribourg (2019)

Posted in Activity by Tomas Rubin on June 21, 2019

The colleagues from Fribourg organised Doctoral day in June 2019 for the PhD students of the universities in western Switzerland (Conférence universitaire de Suisse occidentale, CUSO). I gave a talk on my core research around Sparsely observed functional time series.