Tree Generation Methods Comparison in GAP Problems with Low Quality Data

Data gathered from real world processes include several undesired effects, like the noise in the process, the bias of the sensors and the presence of hysteresis, among other undesirable uncertainty sources. Learning models using the so called Low Quality Data (LQD) is a difficult task which has been barely studied. In a previous study, a method for learning white box models in the presence of LQD that makes use of Multi Objective Simulated Annealing hybridized with genetic operators method for learning models was proposed. This research studies the role of the tree generation methods when learning LQD. The results of this study show the relevance of the role of tree generation methods in the obtained results.