Design of Accurate Classifiers With a Compact Fuzzy-Rule Base Using an Evolutionary Scatter Partition of Feature Space

Abstract

Abstract-An evolutionary approach to designing accurate classifiers with a compact fuzzy-rule base using a scatter partition of feature space is proposed, in which all the elements of the fuzzy classifier design problem have been moved in parameters of a complex optimization problem. An intelligent genetic algorithm (IGA) is used to effectively solve the design problem of fuzzy classifiers with many tuning parameters. The merits of the proposed method are threefold: 1) the proposed method has high search ability to efficiently find fuzzy rule-based systems with high fitness values, 2) obtained fuzzy rules have high interpretability, and 3) obtained compact classifiers have high classification accuracy on unseen test patterns. The sensitivity of control parameters of the proposed method is empirically analyzed to show the robustness of the IGA-based method. The performance comparison and statistical analysis of experimental results using ten-fold cross validation show that the IGA-based method without heuristics is efficient in designing accurate and compact fuzzy classifiers using 11 well-known data sets with numerical attribute values.Index Terms-Fuzzy classifier, intelligent genetic algorithm, orthogonal experimental design, scatter partition.

A general framework for designing a fuzzy rule-based classifier

“…In Table 2, we also provide the average test data set classification accuracy obtained by other authors for the same public data sets in recent studies. In [65][66][67], evolutionary techniques have also been used to design the classifiers. As can be seen from Table 2, the proposed technique outperformed the other approaches on all the three data sets.…”

V-ANFIS for Dealing with Visual Uncertainty for Force Estimation in Robotic Surgery

“…Here, i and j stand for the i−neuron and j−rule, while S i j are the regions of the input space partitioning. In this work, we use scatter partition since it is well-known by its effective performance in comparison with other approaches [37]. It also avoids both the dimensionality problem and data interpretability degradation.…”

Evolutionary Algorithm Parameter Tuning with Sensitivity Analysis

“…De Castro et al in [7] studied the sensitivity of 3 parameters (number of antibodies, number of generated clones and amount antibodies to be replaced) of their Clonal Selection Algorithm (CLONEALG). Similarly, Ho et al in [13] have analyzed the sensitivity of parameters of their Intelligent Genetic Algorithm (IGA), including mutation and crossover probabilities. In [20] Min et al analyze the sensitivity of the population size and the termination condition (maximum number of generations) of a standard GA on a reverse logistics network problem.…”