D.S. Kalana Mendis, Asoka S. Karunananda, U. Samarathunga
This paper presents a real world application where intelligent techniques have given promising results when statistical methods fail. Here we have considered the domain of Ayurvedic classification of individuals. Ayurvedic classification on individuals is based on analysis of a questionnaire. The questionnaire has been used over many years without any research into its improvement. So, classification using this method is still vague and subjective. An intelligent hybrid system has been developed to recognize human constituents. The system consists of modules using two intelligent techniques, namely, Fuzzy logic and Expert system.
The statistical analysis on questionnaire of pilot study shows that principal component (PC) are not significant to consider. However, according to intelligent system developed, there is a significant difference between what we conclude with PC analysis and without PC analysis, provided that intelligent technique of Fuzzy logic is used. This is an interesting result that shows limitation of statistical techniques and how intelligent systems can be used for improving decision making.