Machine learning

This research area is concerned with the development of a Knowledge Acquisition Tool for second-generation Knowledge Based Systems. In the sense of the above mentioned purpose, there is an automatic rule generator being developed, using induction, conceptual-clustering algorithms, potential relevance. The studies about conceptual-clustering algorithms are performed in collaboration with the Institut d'Investigació en Intel.ligencia Artificial (IIIA) and the Universitat Rovira i Virgili (URV). These tools will be integrated in the MILORDII system developed at IIIA. 

We are also studying the Relevance Generating Functions for classification. The main objective of this study is to unify the Relevance Measures proposed by Quinlan, Lopez de Mantaras and others. 

There are three main applications under development: 

  •      Mental-pathologies classification, in collaboration with the Medicine Department of the Universitat Autonoma de Barcelona. 
  •      Sponge classification, in collaboration with the Marine Ecology Department (CEAB). 
  •     Wastewater treatment-plant (WWTP) management, in collaboration with two other research groups: 
    •        The Chemical and Environmental Engineering Laboratory (Laboratori d'Enginyeria Química i Ambiental - LEQUIA) of the University of Girona (Universitat de Girona - UdG
    •        A group, working on WWTPs, of the Chemical Engineering Department (Departament d'Enginyeria Química - EQ) of the Autonomous University of Barcelona (Universitat Autònoma de Barcelona - UAB). 
Participants: J. Béjar (UPC), U. Cortés (UPC), J.M. Gimeno, E. Plaza (IIIA-CSIC), David Riaño (URV), C. Sierra (IIIA-CSIC), V.Torra (URV).