Abstract
Clinical diagnosis by neural networks. Demonstrates the use of neural network expert systems for diagnostic purposes in the field of clinical psychology. Data from 68 patients (17 of them diagnosed as depressive by a clinical psychologists' team, 17 neurotic, 17 squizofrenic and 17 "without pathologies" patients) were collected in a 169 ítems' questionaire. Then a multilayer back-propagation neural network was trained with these data. A proportion of 0.917 new patients were correctly diagnosed by the net. These results show how neural netwoks can be applied as an useful tool in the clinical diagnostic.
