Early Detection of Autism Spectrum Disorder in Children using Machine Learning Techniques and Densenet Model

Autism Spectrum Disorder (ASD) is a neuro ailment that can have a lifelong impact on a Children speech, language development, cognitive function, and social skills. It affects about 1% of people worldwide, and its symptoms typically appear during the developmental stages, or the first two years following birth. The primary causes of ASD are either genetic or environmental, although early detection and treatment can improve the condition. To diagnose ASD, clinical standardized tests are the only techniques performed currently. This leads to a significant rise in medical expenses in addition to requiring more time for diagnostics. The traditional approaches are being enhanced with machine learning techniques to improve the precision and time needed for diagnosis. According to the outcomes, we have constructed predictive models such as KNN, Random Forest Classifier (RFC), Support Vector Machines (SVM), and Logistic Regression (LR). on our dataset. Densenet-121 model being applied to detect the child’s facial expressions and their emotions. Our paper's primary goal is to find out if at the initial stages, the child is at risk for ASD which will assist in the diagnosing process. based on our analysis, the best technique for the dataset we selected is logistic regression.