Optimising weights for heterogeneous ensemble of classifiers with differential evolution

摘要: The classification performance of a weighted voting ensemble classifiers largely depends on the proper weight chosen for each base classifier's vote. In this paper, we propose use Differential Evolution algorithm adjustment voting-weights used in heterogeneous (HEoC). We average Matthews Correlation Coefficient (MCC), calculated over 10-fold cross-validation, as quality measure an ensemble. applied vanilla DE to maximise MCC score training dataset. optimises classifiers' weights order attain better generalisation testing datasets. Experiments were performed using 10 binary-class datasets taken from UCI-Machine Learning Repository. results show consistent and superior constructed ensembles when compared with other well-known classifiers.