Simultaneous Differential Diagnoses Basing on MMPI Inventory Using Neural Networks and Decision Trees Methods

Problem description: Our research concerns psychometric diagnoses basing on the Minnesota Multiphasic Personality Inventory (MMPI) test [1, 3, 4]. The goal of MMPI test is to count the personality–psychometric dimensions which help in differential diagnosis of given person and to assess psychotherapy results. Finally given person should be assigned to one (or two if given case is placed on a border) of personality types. Unfortunately even after the psychometric (MMPI) dimensions are obtained the diagnosis is still so complex for many cases. Our goal is to build a model using neural networks and machine learning technics which could be able to diagnose nosological type and will reduct above difficulties of MMPI test. It is also important to evaluate the influence of each scale in a given patient diagnosis, present alternative classes and estimate the accuracy of such diagnosis.