Multiple objective evolutionary algorithm for temporal rule extraction

Autonomous linguistic rule extraction involving temporal dynamics is an application of growing interest for its relevance to both decision support systems and fuzzy controllers. In the presented work, rules are extracted by and evaluated using three qualitative metrics (i.e., rule inaccuracies, antecedent/consequent mismatches and lack of supporting evidence in the dataset) based on their representation on the truth space diagram. Each metric is then treated as an independent competing optimization goal and multiple objective evolutionary algorithm is used to obtain a set of optimal non-dominant rules. Each representing the best rule observed from the given dataset. Novel techniques for data preprocessing and rule set postprocessing are designed. Data collected from an experimental hot and cold water mixer is used to validate the proposed procedure.