Detection of Autism Spectrum Disorder in Eye Tracking using Hybrid Machine Learning and Deep Learning Technique

One of the common diseases that are being left unnoticed among children and adults is Autism Spectrum Disorder (ASD). Significant research has been carried out to understand the nature of this disorder with the help of eye movements. Lack of eye contact is considered a prime factor in autistic traits. The eye gaze information was the hallmark of autism, and it was used to identify the disorder at an early stage. Machine learning techniques are widely applied to develop the model, which helps in the detection of autism at the early stage of children’s development. The existing model faces great difficulty investigating eye-tracking data from face-to-face conversations. The proposed model tends to overcome these problems and provides high accuracy by using Hybrid K-Means and Fuzzy C-Means through cluster head and Modified Random Forest, and Multi-Layer Perceptron (M-RFMLP) is applied to classify the autistic traits in individuals.