Output details

A fuzzy numeric inference strategy for classification and regression problems

<22>Presents a novel strategy for selection, optimisation and application of a fuzzy numeric inference strategy to both classification and regression problems. The main significance is that the resultant fuzzy trees are easily interpreted, provide a transparent decision making process, and minimize the effect of misclassification of continuous attributes close to decision thresholds. Empirical results presented on known machine learning datasets demonstrate creation of robust, flexible fuzzy trees with an improved performance compared to that of traditional crisp trees. Research informs work on fuzzy decision trees in industry, e.g., Attar Software Ltd (http://www.xpertrule.com/pages/fuzzy.htm) via established collaborations with Crockett.