Dynamic Ensemble Selection and Instantaneous Pruning for Regression Used in Signal Calibration
A dynamic method of selecting a pruned ensemble of predictors for regression problems is described. The proposed method enhances the prediction accuracy and generalization ability of pruning methods that change the order in which ensemble members are combined. Ordering heuristics attempt to combine accurate yet complementary regressors. The proposed method enhances the performance by modifying the order of aggregation through distributing the regressor selection over the entire dataset. This paper compares four static ensemble pruning approaches with the proposed dynamic method. The experimental comparison is made using MLP regressors on benchmark datasets and on an industrial application of radio frequency source calibration.
KeywordsEnsemble Pruning Dynamic Ensemble Selection Ensemble Methods Calibration
Unable to display preview. Download preview PDF.
- 2.Windeatt, T., Zor, C.: Ensemble Pruning Using Spectral Coefficients. IEEE Trans. Neural Network. Learning Syst. 24(4), 673–678 (2013)Google Scholar
- 9.Cavalin, P.R., Sabourin, R., Suen, C.Y.: Dynamic Selection Approaches for Multiple Classifier Systems. In: Formal Aspects of Cognitive Processes. LNCS, vol. 22 (3-4), pp. 673–688. Springer (2013)Google Scholar
- 11.Shen, Z.-Q., Kong, F.-S.: Dynamically Weighted Ensemble Neural Networks for Regression Problems. Machine Learning and Cybernetics, 3492–3496 (2004)Google Scholar
- 12.Mendonca, M., Da Silva, I.N., Castanho, J.E.C.: Camera Calibration Using Neural Networks. Journal of WSCG 10(1-3), POS61–POS68 (2002)Google Scholar
- 13.Khan, S.A., Shahani, D.T., Agarwala, A.K.: Sensor calibration and compensation using artificial neural network. ISA Transactions 43(3) (2003)Google Scholar
- 14.Wang, D.-S., Liu, X.-G., Xu, X.-H.: Calibration of Arc-Welding Robot by Neural Network. Fourth International Conference on Machine Learning and Cybernetics, Guangzhou (2005)Google Scholar