An Effective Ensemble-based Machine Learning Classifier for ASD with Feature Selection

K. Vijayalakshmi, Dr.M. Vinayakamurthy, Dr. Anuradha and Saiprasad

Nowadays to revolutionize medical diagnosis, machine learning techniques plays very important role as it has the potential to deal with prediction accuracy. Autism spectrum disorder (ASD) is growing at a very fast rate, characterized by notable phenotypic heterogeneity which becomes a growing concern in medical research. If the detection of ASD is done early, then it helps to notify the risk level and provides the effective remedial solutions by achieving accuracy in prediction. But for high dimensional dataset, it is very crucial for successful and promising diagnosis and treatment. Hence in this paper, relevant features from ASD Screening Data for Adult Autism dataset are selected by using Relief algorithm. It is further applied on different classification algorithms. Finally, the performance analysis proven that the best classifier is Random forest according to its accuracy and error rate measures.

Volume 11 | 04-Special Issue

Pages: 1667-1672