Interpretability issues in evolutionary multi-objective fuzzy knowledge base systems

Abstract

Interpretability and accuracy are two conflicting features of any Fuzzy Knowledge Based System during its design and implementation, this conflicting nature leads to Interpretability-Accuracy Trade-Off. Secondly, the assessment of interpretability is another important problem for researchers. Several indexes and methods have been proposed for assessing interpretability but this issue remains an open problem because of its subjective nature. This paper discusses two research issues, interpretability assessment and interpretability-accuracy trade-off in Fuzzy Knowledge Base System design using Evolutionary Multiobjective Optimization by proposing taxonomy for studying and evaluating the interpretability. © 2013 Springer.