Autism Spectrum Disorder Prediction using Ma...

Autism Spectrum Disorder Prediction using Machine Learning

Author Name : M. Priyadharshini, Dr. R. Sri Devi

DOI: https://doi.org/10.56025/IJARESM.2025.130425283

ABSTRACT Autism Spectrum Disorder (ASD) is a Neuro developmental condition affecting social interaction, communication, and learning. Diagnosing ASD is time-consuming and costly, especially with traditional methods. Early detection can help mitigate the progression and long-term impact of the condition. This project aims to use machine learning techniques to predict autism spectrum disorder (ASD) using various features such as demographic, behavioral, and diagnostic information. The features of predictions are using ID , Gender , Age , Ethnicity , Jaundice, Class /ASD , Autism , behavioral scores (e.g., responses to A1_Score to A10_Score). Employing various preprocessing techniques and classification algorithms, the study achieves high accuracy in autism prediction, highlighting the potential of automated tools in clinical diagnostics. The process includes data loading, understanding, and preprocessing to clean, normalize, and prepare the dataset for model training. Through Exploratory Data Analysis (EDA), the relationships between features and the target variable are examined. In the presented work Decision Trees, Random Forests, and XG Boost, are trained and evaluated using performance metrics like accuracy, precision, recall, and AUC-ROC. Hyper parameter tuning is applied to find the best set of parameters for optimal model performance. The final model selected is based on its ability to generalize well to unseen data, providing insights that can help in the early diagnosis and intervention of ASD.