MAIN DETERMINANT FACTORS AFFECTING MULTIV ARIATE CLASSIFICATION SUCCESS RATIO USING POTENTIAL CURVES

The usage of potential curves constitute a simple, fast and effective way for classifying new, unknown samples into a predefined set of grou ps. In this work a discussion is presented regarding the two most impo rtant factors affecting the classification success ratio: the analytic al parameters employed to chemically characterize the samples and the relative distribution of the predefined groups. As a practical example , two sets of tea samples were classified into different categories.