MEGA: Predicting the best classifier combination using meta-learning and a genetic algorithm

Classifier combination through ensemble systems is one of the most effective approaches to improve accuracy classification systems. Ensemble are generally used combine classifiers; However, selecting best individual classifiers a challenging task. In this paper, we propose an efficient assembling method that employs both meta-learning and genetic algorithm for selection classifiers. Our called MEGA, standing using MEta-learning Genetic Algorithm recommendation. MEGA has three main components: Training, Model Interpretation Testing. The Training component extracts meta-features each training dataset uses discover classifier combination. interprets relationships between priori multi-label decision tree algorithms. Finally, Testing weighted k-nearest-neighbors predict unseen datasets. We present extensive experimental results demonstrate performance MEGA. achieves superior in comparison other methods and, importantly, able find novel interpretable rules can be select dataset.