Prior Prediction and Management of Autism in Child Through Behavioral Analysis Using Machine Learning Approach

Abstract

AbstractAutism is a neuro-developmental disorder that involves a delay in those areas of development and early diagnosis can reduce these problems. Autism in Child prediction will be more beneficial if possible to predict autism in the preliminary stage and this research will be helpful for physicians, parents, and doctors before the formal screening test. On the other hand, existing screening tools for early prediction of autism involve long-term observation and in-depth evaluation by the experts are time consuming and expensive. Therefore, Machine Learning (ML) approach may help the professionals to diagnose of autism in its early stages by the pretrained model with existing data. In this paper, we have analyze the machine learning algorithms to determine a set of conditions that together prove to be predictive of autism disorder. Moreover, this paper proposed a autism management system using knowledge fusion of gathered features from parents monitoring mobile apps. This research paper analyzed about 65 attributes of child under the different behavioral categories to improve the accuracy of the prediction ML model. Additionally, we have adopted Decision Tree (DT), Multinomial Naive Bayes (MNB), Random Forest (RF), Adaboost, Multilayer Perceptron (MLP), K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Logistic Regression (LR) is applied on a dataset of 500 children to compare with a weighted voting approach to achieve higher accuracy of 97%, where the Weighted Voting Classifier (WVC) outperforms than the other state-of-the-art classifiers.KeywordsAutismMachine learningWeighted voting classifierDisorderASD