Human Random Generation as a Probe to Mental Condition

Abstract

Human random generation test (HRG) is expected to be used for detecting a mental disease in a handy, noninvasive way. In order to implement such function in a simple apparatus, we need a quick way to compute an appropriate measure, or to recognize patterns to be used for diagnosis. We examine the effectiveness of hidden Markov model (HMM) for HRG by using two sets of data which we took in 1997 and 2003 from 13 students in total. Our result implies that HRG offers information enough to identify group of people but not enough to recognize individuals by means of HMM.