ADHD/Autism Disorders Clustering Method Using Korean Morphology Analyzer with Observed Behaviors Data

Abstract

In this paper, we propose a method to cluster Attention Deficit/Hyperactivity Disorder (ADHD) and autism disorder using observed behaviors data. We utilized a Korean morphology analyzer to analysis features of the observed behaviors data which was recorded by text type. The proposed method aims to support an assistance tool to a special education and rehabilitation teachers to reduce the educating efforts. To show the efficiency of the proposed method, we used the historical text-typed records accumulated in the Research Institute of Special Education and Rehabilitation Science (RISPERS), Daegu University. The data had been recorded for 5 years. As a result of simulation, we confirmed that the proposed method achieved the clustering accuracy of 60% by choosing and evaluating the random result sample of 200 children of disorders.