Dušan Sovilj
M.Sc. (Tech.), Researcher
- Office:
-
Room A344 in Computer Science Building,
Konemiehentie 2, Otaniemi campus area, Espoo - Postal Address:
-
Aalto University School of Science,
Department of Information and Computer Science,
P.O. Box 15400, FI-00076 Aalto, Finland - Telephone:
- +358 45 327 33 11 (personal)
- Email:
- firstname.lastname [at} aalto.fi
Description
Ph.D. student in TSPCi (now EIML) group started in Sep 2009. I received my B.Sc. in Computer Science at University of Novi Sad (Serbia) in 2006, and my M.Sc. in Technology at HUT in 2009.
Research
Time Series Prediction
Variable Selection for Regression
Environmental Modelling
Abstract
Topic of my Master thesis was variable selection using Delta test in context of selection, scaling and projection, used as a preprocessing step before long-term time series prediction. Currently, my Ph.D. research topic is on developing and applying machine learning tools for Baltic Sea data sets to be used in domain of environmental modeling.
Had a research visit to the Leibniz Institute for Baltic Sea Research Warnemünde in Rostock in June/July 2010 for some intensive biology - machine learning collaboration. Second visit to Rostock in July/August 2011 and again, the same institute.
June 2011, change of working place from B307 to A344. Hopefully my candy stash will remain intact in the new place.
Publications
List of Publications. Order by: Type | Date
2. Relevance learning for time series inspection
Andrej Gisbrecht, Dušan Sovilj, Barbara Hammer and Amaury Lendasse.
In ESANN'12. April, 2012, accepted for publication.
1. Climate induced changes in benthic macrofauna -- a non-linear model approach
Karin Junker, Dušan Sovilj, Ingrid Kröncke and Joachim Dippner.
In Journal of Marine Systems, volume 96-97, pages 90-94. August, 2012.
2. Variable Selection in a GPU Cluster Using Delta Test
Alberto Guillén, Mark van Heeswijk, Dušan Sovilj, Maribel García Arenas, Luis Javier Herrera, Hector Pomares and Ignacio Rojas.
In IWANN (1), pages 393-400. 2011.
1. Multistart Strategy using Delta Test for Variable Selection
Dušan Sovilj.
In ICANN 2011, volume 6792, pages 413--420. June, 2011.
3. Extending Self-Organizing Maps with Uncertainty Information of Probabilistic PCA
Dušan Sovilj, Tapani Raiko and Erkki Oja.
In IJCNN, pages 1915--1921. July 18-23, 2010.
2. Approximate k-NN Delta Test Minimization Method using Genetic Algorithms: Application to Time Series
Fernando Mateo, Dušan Sovilj and Rafael Gadea.
In Neurocomputing, volume 73, pages 2017-2029. June, 2010.
1. OPELM and OPKNN in long-term prediction of time series using projected input data
Dušan Sovilj, Antti Sorjamaa, Qi Yu, Yoan Miche and Eric Séverin.
In Neurocomputing, volume 73, pages 1976-1986. June, 2010.
2. RCGA-S/RCGA-SP Methods to Minimize the Delta Test for Regression Tasks
Fernando Mateo, Dušan Sovilj, Rafael Gadea and Amaury Lendasse.
In IWANN 2009, volume 5517, pages 359-366. June 10-12, 2009.
1. Fast Variable Selection using Delta Test
Dušan Sovilj.
September, 2009.
3. New Methodologies Based on Delta Test for Variable Selection in Regression Problems
Alberto Guillén, Dušan Sovilj, Fernando Mateo, Ignacio Rojas and Amaury Lendasse.
In Workshop on Parallel Architectures and Bioinspired Algorithms. October 25-29, 2008.
2. Tabu Search with Delta Test for Time Series Prediction using OP-KNN
Dušan Sovilj, Antti Sorjamaa and Yoan Miche.
In ESTSP, European Symposium on Time Series Prediction, pages 187-196. September 17-19, 2008.
1. Minimizing the Delta Test for Variable Selection in Regression Problems
Alberto Guillén, Dušan Sovilj, Fernando Mateo, Ignacio Rojas and Amaury Lendasse.
In International Journal of High Performance Systems Architecture, volume 1, pages 269-281. 2008.
Material on this web site is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders, notwithstanding that they have offered their works here electronically. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder. Copyright holders claiming that the material available below is not in accordance with copyright terms and constraints are invited to contact the author by e-mail and ask him to remove the links to specific manuscripts. Most PDF files of journal articles and book chapters contain a non-final version of the manuscript (but still, sufficiently close to it...).