Data Science Group
Head of the research group: Professor SADOK BEN YAHIA, email@example.com
Topics and Competences
The Data Science Group is carrying out research activities towards extracting value from information. standing at the crossroads of some of priority areas, e.g.
- Cyber security: Intrusion /outlier detection
- Digital transformation of the society; healthcare information systems (Adverse events in hospitals, mining of patient trajectory)
- Smart-environment: IOT : Intelligent transportation system, smart city, smart home
Main competences of the research research group are:
- Data Mining
- Big data analytics
- Data seriation /visualization
- Ontology alignment/merging
- Random data modeling
A language-independent system has been created for testing dictionaries of the Wordnet type (there are about 70 of them in the world). It was shown that all dictionaries of this type contain tens of thousands of errors that the new testing system effectively detects. Cooperation with the authors of the Estonian Wordnet, for example, reduced the number of polysemic errors in the Estonian Wordnet by 97%.
An existing algorithm of zero-factor-free determinacy analysis was developed in order to find out how to detect an object’s belonging to a class. It is based on finding closed sets and their generators at the same time. As a bases for describing underlying data three types of rules can be found: class detection rules, (positive) association rules, negative association rules. The usage methodology was developed.
Recent Research Results
An effective copyrighted formal method was created and programmed: “How to find corruptive coalitions in any procurement systems” (L. Võhandu and A. Lohk). An application of text mining to detect synonyms automatically from a web dictionary was created (A. Lohk, M. Tombak, K. Vare).
A monotone systems theory based algorithm for finding equivalent classes was developed (Grete Lind, Rein Kuusik).
- Ross, K.; Lohk, A. (2017). Words, Forms and Phrases in Estonian Folksongs and Hymns. Folklore. Electronic Journal of Folklore, 67, 49-64.10.7592/FEJF.
- Henno, J. Information and interaction. Information Modelling and Knowledge Bases XXVIII / Amsterdam: IOS Press, 2017, 426–449. (Frontiers in Artificial Intelligence and Applications; 292).
- Võhandu, L.; Tamme, T.; Rull, A. (2017). Some formal methods for language ecology. Abstracts: EAAL 16th Annual Conference Language as an Ecosystem, April 20–21, 2017, Tallinn, Estonia. Eesti Rakenduslingvistika Ühing, 10.
|Jaak Henno||Senior Research ScientistSchool of Information Technologies: Department of Software Sciences||ICTfirstname.lastname@example.org||6202317|
|Grete Lind||SpecialistSchool of Information Technologies: Department of Software Sciences||ICTemail@example.com||6202306|
|Ahti Lohk||LecturerSchool of Information Technologies: Department of Software Sciences||ICTfirstname.lastname@example.org||6202318|
|Martin Rebane||Early Stage ResearcherSchool of Information Technologies: Department of Software Sciences||ICTemail@example.com|
|Ants Torim||LecturerSchool of Information Technologies: Department of Software Sciences||ICTfirstname.lastname@example.org||6202306|
|Tarvo Treier||LecturerSchool of Information Technologies: Department of Software Sciences||ICTemail@example.com||6202316|
|Tauno Treier||School of Information Technologies: Department of Software Sciencesfirstname.lastname@example.org|
|Tarmo Veskioja||Research ScientistSchool of Information Technologies: Department of Software Sciences||ICTemail@example.com||6202345|
|Leo Võhandu||ConsultantSchool of Information Technologies: Department of Software Sciences||ICTfirstname.lastname@example.org||6202346|
|Professor Emeritus: School of Information Technologies||ICT-644|