Enhanced of autism spectrum disorder using grey relational analysis and supervised learning for classification

• Main Authors: Mohd. Yunos, Z., Idrus, W., Shamsuddin, S. M., Saadon, M. S. I., M. Yusuf, S.
• Format: Conference or Workshop Item
• Language: English
• Published: 2020
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/93599/1/ZuriahatiMohdYunos2020_EnhancedofAutismSpectrumDisorder.pdf; http://eprints.utm.my/id/eprint/93599/; http://dx.doi.org/10.1088/1757-899X/864/1/012083

Autism spectrum disorder (ASD) is complicated to be diagnosed and many study had shown that machine learning technique have been proven to accurately diagnose ASD. However, there were also some drawbacks in the results obtained and one of it is related to the lower accuracy upon implementation. Thus, a feature selection method namely Grey Relational Analysis (GRA) is proposed to enhance the performance of the machine learning technique in the classification of ASD as it was proven too often produce high accuracy results. GRA is used to select relevant features and rank them from the highest to the lowest. The data used was the adult autism data, which consists of 608 data with 16 features. The machine learning techniques used are Support Vector Machine (SVM) and Artificial Neural Network with Multi Layer Perceptron (ANN-MLP) to classify ASD. From results obtained, the integration of GRA and machine learning techniques have managed to produce a high accuracy of more than 90%. The SVM gave the good accuracy of 98.1%, while ANN produce of 98.36%