DEX method: recent advances and future challenges on developing predictive models from data

In the last decade, decision support and data mining approaches have become indispensable tools for researchers and practitioners in agronomy and related fields. Decision modelling is aimed at developing models, usually through collaboration of decision makers and experts, for suggesting and justifying solutions of the problem at hand. Data mining uses data to find meaningful patterns, most often in the form of predictive models, for a similar purpose. Both types of models can be included in decision support systems for aiding decision makers in recurring decision tasks. This seminar will be given through two 30-minute presentations by two senior researchers from Jožef Stefan Institute, Department of Knowledge Technology, Ljubljana, Slovenia: Marko Bohanec, an expert in decision support and creator of DEX method and DEXi software and Sašo Džeroski, an expert in data mining, leading the development of many data mining tools, and Head of the Department. Marko Bohanec will focus on the method DEX and present some recent advances and plans, for instance inclusion of the option generator method in DEXi. Participants who already know DEXi are kindly asked to pose questions and give comments. Sašo Džeroski will present some methods for mining complex data, in particular methods for multi-target prediction. He will also illustrate their use on several problems from ecology and agriculture.;;