Pattern recognition using multivariate-based fuzzy inference rule reduction on neuro fuzzy system

The pattern recognition of the IRIS data using multivariate-based fuzzy inference rule reduction technique presented. There are numerous data sets to perform or recognize the data themselves with unnecessary results of pattern recognition such as undesired rules or clusters with inappropriate degree precision because of the imprecise and massive data set itself. The proposed data reduction technique reduces the large and imprecise data into the relatively reduced and precise data produced by the consecutive preprocessing factor analysis and subtractive clustering (SUBCLUST) analysis to generate the neuro fuzzy system for pattern recognition. As case study, it is examined for the performance and its accuracy using the proposed technique with IRIS data to recognize pattern from 150 selected flowers from three species and their four different measurements using statistical measurements such as correlation (CORR), total root mean square (TRMS), standard deviation (STD), mean of absolute distance (MAD), and equally weighted index (EWI).