Experiments with Fuzzification of MMPI Profiles for the Classification Task

Abstract The paper is devoted to experiments with fuzzification of MMPI (Minnesota Multiphasic Personality Inventory) profiles for the classification task. MMPI is a standardized psychometric test of adult personality and psychopathology. In our study, we have used a classifier built on the basis of fuzzy decision trees generated by means of the algorithm that uses cumulative information estimations of the initial data proposed by V. Levashenko et al. The main attention is focused on testing a variety of approaches to fuzzification of the MMPI scales that make up the profiles. Both several shapes of membership functions and different ways of divisions of scale ranges are considered.