On Effectively Predicting Autism Spectrum Disorder Using an Ensemble of Classifiers