Preictal phase detection on EEG signals using hybridized machine learning classifiers with a novel feature selection procedure based GAs and ICOMP

•Pre-ictal phase detection for epileptic seizure.•A novel training and feature selection procedure for ML classifiers.•Hybrid training with feature selection based GAs and information criteria.•The robust models in terms of balancing model complexity.•Identifying pre-ictal patterns with high accuracy and lower FP and FN.