Projects
Online electronic platform for exercises in Statistics. Vrije Universiteit Brussel
statisticl modeling of visual patterns wiht applications to digital painting analysis Ghent University
the goal of this project is to develop new statistical models and representations of visual patterns and thereby bring new and important advances in the analysis of digitized paintings. the project aims at both fundamental theoretical contributions in vision science and at creating innovative computer vision methods for analysis and authentication of digitized masterpieces.
Computational challenges in robust statistics KU Leuven
Robust statistics aims at detecting anomalies in data by looking for the pattern followed by the regular measurements. This often leads to complex estimators for which efficient algorithms need to be constructed. The difficulty increases considerably when considering high-dimensional or ultrahigh-dimensional data such as images, videos and omics data. The project aims at developing methods and algorithms that are able to detect outlying cases ...
PhD researcher in Statistics KU Leuven
The thesis is focused on modeling dependencies (e.g. via Copula) and post-selection inference. The main topics include statistical inference, model selection, post-selection inference, and nonparametric statistics. In addition, the research also involves trending topics in modern statistics, such as functional data analysis, statistical learning theory, empirical process theory, concentration of measures, etc.
Scientific and methodological support for the Brussels Institute for Statistics and Analysis (BISA) of perspective.brussels - Lot 4: Demography Vrije Universiteit Brussel
The use of machine learning in combination with classical statistics for better predictive modelling in insurance KU Leuven
The research will focus around the use of data science, classical statistical methods and machine learning techniques within an actuarial framework.
Data Science for Official Statistics. KU Leuven
In general, research will be conducted into the use of data science to produce new and better statistics. The challenges are varied and range from the efficient use of new data sources to the correct use of complex machine learning algorithms and innovative data visualization techniques; identifying new applications in this field and investigating the feasibility of and challenges in new data science opportunities.
LOKSTAT +. Database of local statistics in Belgium (1960-1990). Ghent University
LOKSTAT + is an extension of LOKSTAT, the database of local statistics in Belgium (1800-1970). LOKSTAT (developed at UGent) is a digital database of demographic, socio-economic and political-institutional statistics, available from 1800 for the entire Belgian territory, down to the municipal level. Whereas LOKSTAT was compiled on the basis of paper archives from the period 1800-1970 (the pre-computer era), LOKSTAT + focuses on the period ...
Essays on multivariate statistics in finance KU Leuven
The research concerns multivariate statistical methods to model and estimate financial asset prices. In one strand the research will be concerned with estimating systemic risk and extreme dependencies between asset prices in a more flexible fashion. In another strand the effects of outliers in asset prices and pricing factors on the estimated model structure will be considered.