Autism’s Disease Prediction Using Deep Learning

Getting healthcare is a necessity for human survival. The phrase (autism disease) is wide and includes a range of symptoms that are used to diagnose it. Early diagnose disease assessment techniques helped determine the appropriate course of events for high- peril people, hence reducing their peril. The main goal is to protect people by identifying abnormalities in behavior. Autism’s disease prediction are developed. Which Early diagnose disease. With this, the model needs improvement. For that purpose, in this paper we present a hybrid model utilizing Convolutional neural network and Bidirectional Long Short-Term Memory (CNN and Bi-LSTM) that uses deep learning techniques to predict whether or not an individual has autis. Using data processing techniques, we address the problems of missing and unbalanced data in the publicly available Autism Screening on Adults dataset. For feature selection, we have employed a more tree classifier, and for classification, CNN-BiLSTM. The dataset gathered from Kaggle has been the subject of experiments. The diagnosis model’s performance is evaluated using methodologies including classification, accuracy, precision, recall, and f1-score. In order to propose an accurate model for diagnosing autism spectrum disorder, a study is carried out here utilizing the recommended approach and their performance analysis. With an examination of several current techniques, the results of hybrid model achieved an accuracy of 97%