A Multi-Classifier-Based Recommender System for Early Autism Spectrum Disorder Detection using Machine Learning
Anita Shinde, Dipti D. Patil
Abstract
Efficient and effective medical diagnostic systems are needed for Autism Spectrum Disorder (ASD) detection and treatment. Healthcare specialists usually generate extensive remarks on patient behavioural assessment, which is time-consuming to process and record. Early detection of ASD means quality life with the help of appropriate treatment and care. Machine learning models can be utilized to investigate the feasibility of identifying the stated features and evaluating the presence or absence of autism. This study develops a recommender model with multi-classifiers to enhance precision in the prediction of ASD. Random Forest (RF), Naïve Bayes (NB), Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM), and Artificial Neural Network (ANN) are proposed to assess the model’s performance. We show that Decision Trees and Random Forests exhibit improved performance if analyzed with other algorithms regarding accuracy, precision, recall, and F1-score as evaluation metrics. In the future, work can be extended with different datasets and can focus on speedup the process of generating recommendations.