Kvalitativno modeliranje na osnovi podatkov (Slovene)

Developed methods represent a pioneering work in the field of qualitative modeling. The work is especially interesting because of innovative connecting of techniques from different fields - topology, symbolic computation, machine learning, numerical analysis, probability and statistics. The core of the project, however, belongs to the field of artificial intelligence. Due to the limited time and resources available we did not expect the project to develop qualitative modeling to the same level of maturity as that of classification and regression learning which have been developing for half a century by a large community. We however believe that we provided a good basis for its future development. To project's results are useful in many other areas of science that rely on machine learning and data mining. These include all sciences that derive hypotheses from empirical data, most notably modern genetics and medicine. Developed algorithms will be, for instance, useful in analysis of dependencies between genes in genetic network, which can be used in modeling and curing diseases on genetic level. Other examples of scientific fields that rely heavily on drawing conclusions from experimental data are social sciences, psychology and economy and also most other areas of modern science.

The work of Slovenian researchers in AI has always represented the state­of­the­art in the field and was also very successful with regard to obtaining EU­-funded research grants. Our work in the unexplored area of qualitative modeling will help it to maintain this position.