An analysis of a multidimensional dataset of an epidemic study using soft computing tools - a pilot study

Two contrasting approaches toward an epidemic study were illustrated as a pilot study; the regression analysis which is rather conventional methodology used in the past/present epidemic studies, and the other is the classifier analysis which is in the soft computing toolbox. The dataset we used for this study is obtained from a part of a cohort study which principally focused on a fatigue syndrome of the elementary and junior high school educates. In the classifier analysis we employed a major supervised machine-learning algorithm, K-Nearest Neighbour (K-NN), coupled with Principal Component Analysis (PCA). As a result, the performance that was found by cross validation method in the classifier analysis provides better results than that of the regression analysis. Finally we discussed the availability of both analyses with referring the technical and conceptual limitation of both approaches.