9) University of Cyprus

Scientific Staff

Prof. Erricos J. Kontoghiorghes
Prof. Chris Charalambous
Dr. Cristian Gatu
Dr. Petko Yanev
Dr. Paolo Foschi

Erricos John Kontoghiorghes has received a B.Sc. in computer science and statistics and a Ph.D. in computer science (focus computational statistics) from Queen Mary College, University of London. He is faculty member of the Department of Public and Business Administration, University of Cyprus, and holds a visiting professorship at Birkbeck College, University of London.

Expertise of the Team

The Department of Public and Business Administration, University of Cyprus has a long tradition in the area of Computational Economics and Finance. It hosts the “Hermes Center of Excellence”, which has gained a reputation as a leading research institute in the areas of computational finance and economics. Research is pursued in computational statistics, management science, finance and economics. The computational statistics aspects are pursued in close collaboration with the ERCIM (European Research Consortium in Informatics and Mathematics) working group on “Matrix Computations and Statistics” and the research group on “Matrix Computations and Statistics” at the University of Neuchâtel, Switzerland (Foschi, Gatu and Yanev). This research focuses in the development of computational methods for the numerical estimation of large-scale linear econometric models (Foschi and Kontoghiorghes, 2003) and computationally intensive methods for statistical model selection (Gatu and Kontoghiorghes, 2006). In addition, parallel algorithms are designed and implemented (Yanev and Kontoghiorghes, 2005). Finally, the effectiveness of the algorithms and implementations are verified in the context of different applications.


  • P. Foschi and E.J. Kontoghiorghes: “Estimating SUR Models with VAR Disturbances”, Journal of Economic Dynamics & Control, 2003.
  • C. Gatu and E.J. Kontoghiorghes: “A Branch and Bound Algorithm for Computing the Best Subset Regression Models”, Journal of Computational and Graphical Statistics, 2006.
  • P. Yanev and E.J. Kontoghiorghes: “Efficient Algorithms for Estimating the General Linear Model”, Parallel Computing, 2006.

Expertise in Training Young Scientists

Erricos Kontoghiorghes and Chris Charalambous have supervised over 10 Ph.D. theses. Currently they are supervising 4 Ph.D. students. Erricos Kontoghiorghes has organized various advance postgraduate research seminars in computational statistics. These seminars have been funded by the Swiss National Foundation for Research.

Links within the Network

Kontoghiorghes, Gilli and Winker have been the investigators of various joint projects of the Swiss National Science Foundation. Currently they are involved in the design of model selection algorithms for vector autoregressive models. Within the framework of a research project on model selection, Gilli and Kontoghiorghes co-supervise a Ph.D. student (Marc Hofmann) and a research associate (Cristian Gatu). In addition they collaborate through various professional activities. Kontoghiorghes and Gilli are co-chairs of the 12th International Conference on “Computing in Economics and Finance” (Cyprus, 2006) and co-organizers of the International Conference on “Computational Management Science” (Geneva, 2007). Further collaboration exists through the editorial work for “Computational Statistics and Data Analysis”: Kontoghiorghes, Gilli and Winker are editors and associate editors of this journal.

Role of the Research Team

The research team will concentrate its contributions on the development of statistical techniques for variable selection in linear and nonlinear models, numerical and computational methods for the solution of the normal equations, and the elaboration and implementation of parallel algorithms. In particular, a branch-and-bound strategy for subset selection of vector autoregressive models will be developed. In addition, optimization heuristics will be considered in order to facilitate the investigation of large scale models. The design of numerically efficient methods for estimating linear econometric models will be pursued. The estimation methods will be based on numerically stable tools such as orthogonal transformations. The parallel implementation of the various strategies on parallel architectures will be considered. The new strategies will be employed to tackle real world problems. The researchers will also contribute to the training and transfer of knowledge activities linked to their specific expertise.