Rytis Bagdziunas

 Profile

I am currently a graduate student in econometrics at the Université catholique de Louvain. As of 2012, I have been awarded Aspirant FNRS scholarship by Communauté Française. It was renewed in 2014 and expires in October 2016.

Since 2012, I have been member of Service d'Analyse Economique team at IRES, UCL. My primary responsibility within the team is analysing and forecasting inflation in Belgium. Reports to which I have contributed can be accessed here.

Education

  • 2011-present: Ph.D. in econometrics, CORE, Université catholique de Louvain
    • Title: Estimation of high dimensional models with risk constraints.
    • Supervisor: Sébastien Van Bellegem
    • Project details: I aim to analyze the issue of endogeneity in the context of high-dimensional data, notably where data exhibits functional smoothness. In particular, we consider functional linear regression models and various regularization techniques for their estimation such as dimension reduction with PCA and Galerkin methods as well as Tikhonov (or L2) regularization.
  • 2009-2011: Master in Mathematics (magna cum laude), Université catholique de Louvain
    • Title: Lefschetz fixed point theorem and its applications
    • Supervisor: Pascal Lambrechts
  • 2008-2010: Research Master in Econometrics (magna cum laude), Université catholique de Louvain
    • Title: Complex dynamics in evolutionary game theory
    • Supervisor: Jean-François Mertens
  • 2005-2008: Bachelor in economics and management (magna cum laude), Université catholique de Louvain

 Work experience

  • 2009-present: Teaching assistant for various university courses. I worked full time during academic year 2011-2012. On the whole, I assisted for microeonomics, (applied) econometrics as well as advanced econometrics courses.
  • 2005-2008: Summer jobs at the Bank of New York Mellon in Brussels

Publications

  • Articles in progress:
    • Estimation of functional instrumental linear regression with Galerkin methods.
    • Statistical inference on instrumental functional regression by Tikhonov regularization.
  • R packages:
    • dynfactoR package which facilitates estimation of state-space based dynamic factor models
    • functionalIV package implementing various estimators for functional linear regression models. This package is part of my Ph.D. program. Both R packages are already usable, though still not tested and not mature enough. Hence, they're both alpha releases.

 Presentations

  • 2015 July: useR!2015 conference in Aalborg, Denmark
  • 2014 November: Belgian Statistical Society meeting in Louvain-la-Neuve, Belgium
  • 2014 July: International Vilnius Conference on Probability Theory and Mathematical Statistics, Lithuania
  • Annual presentations at Doctoral Workshop organized by UCL, UNamur and FUSL

 Programming languages and technical profficiency

  • R
  • SQL, notably SQLite and PostgreSQL with PostGIS
  • Python
  • Git
  • Linux and shell scripting

I am familiar with X13-ARIMA-SEATS statistical adjustment packages as well as modern data exchange protocols, notably SDMX-ML. I am also familiar with Apache Spark basics, notably via PySpark, having followed introductory and scalable machine learning courses on edX.

 Human languages

  • Lithuanian (mother tongue)
  • French (fluent)
  • English (fluent)
  • German (basic)
  • Russian (basic)