Semi-automatic learning with quantitative and qualitative features
Joaquím Comas, Ignasi Rodríguez‐Roda, Luigi Ceccaroni, Miquel Sànchez–Marrè · 2002
Part of artificial-intelligence research is highly concerned with the efficient manipulation of environmental databases. Nowadays there exist several knowledge-based systems, with specialized problem-solving expertise in a large number of areas. And, in order to function, they usually need a leaming module for preliminary knowledge extraction, characterization of the particular subject area, obtaining hints about the problems within that subject area and getting clues about the possible solutions of those problems. In this paper a hybrid leaming system for the wastewater treatment process is described, which includes the participation of human process-experts interacting with two clustering techniques applied to measured descriptors. As a case study, a data set from a wastewater treatment plant (WWTP) was chosen, including 65 quantitative and qualitative features. In the case study, the obtained classes reflect the dynamics of the involved environmental system and are used to identify the state of the WWTP. The final aim is comparing the conclusions obtained from the analysis of only quantitative data with the conclusions achieved after the addition to the analysis of qualitative data and results of microscopic observation.