Predictive Model for Autistic Spectrum Disorder with Neural Network Application

Abstract

The rise in the number of Autistic Spectrum Disorder (ASD) cases across the world reveals that there is a pressing need to develop and implement effective screening methods. The main concern of this study is to formulate a predictive model to detect ASD using data mining technique and neural network. Data mining technique was used to analyze the identified instances using WEKA. A total of 10261 sample data were utilized in this study. Eighty percent of the data was used for training the neural network and twenty percent of the data was used for testing. The classification whether the child has autism or not based on the given patterns was carried out using backpropagation neural network. The calculation of the system resulted in up to 98.37% accuracy. It can be concluded that the backpropagation neural network was able to effectively detect ASD.