MEDICAL DATA PREPARATION WITH AUGMENTATION TECHNIQUES FOR DETECTION OF ASPERGER SYNDROME

Abstract

Asperger syndrome starts to appear in childhood and continues to keep going on into adolescence and adulthood. Propelled by the rise in the use of machine learning techniques in the research dimensions of medical diagnosis, this paper there is an attempt to explore the possibility to use VGG16, Mobilenet v2, Densenet-121, Resnet-51, Inceptionv3, and Convolution Neural Network for predicting A novel data-set is created with ASPERGER SYNDROME individuals of a toddler, adolescent, and adult age groups to evaluate the model. The first data set related to ASPERGER SYNDROME screening in children has 292 instances and 21 attributes. Second data-set related to ASPERGER SYNDROME screening. Adult subjects contain a total of 704 instances and 21 attributes. The third data-set related to ASPERGER SYNDROME screening in Adolescent subjects comprises 104 instances and 21 attributes. ACGAN is applied to increase the data set as there is an imbalance of data between healthy individuals and healthy individuals. After applying various deep learning architectures results strongly suggest that CNN-based prediction models work better on increased data sets with higher accuracy of 99.53, 98.30, and 96.88 % ASPERGER SYNDROME Screening in Data for Adults, Children, and Adolescents respectively.