在分類任務中，集成學習透過整合多個基本模型的預測結果，有效提升整體預測效能，為了進一步解決傳統集成學習在記憶體上的龐大耗損與運算負擔的問題，集成挑選逐漸成為研究焦點，期望在縮減集成規模的同時，仍能維持甚至提升預測效能。集成挑選中，最佳化方法被廣泛應用，過往研究顯示基本模型的正確率與模型預測間的相關性為影響集成效能的關鍵因素，多數相關研究皆將此二者納入多目標範疇，建構最佳化模型。隨著隨機生成基本模型方法的提出，能大幅降低基本模型間預測的相關性，不再需要將此因素納入考量，基於此，本研究提出一個以正確率為單一目標的最佳化集成挑選模型，期望從候選模型中，挑選出最佳的組合進行集成，研究也將過往研究僅適用於二類別分類的最佳化模型框架重新設計，擴展應用範圍至現實中更常見的多類別分類任務。
實驗結果顯示，本研究所提出之最佳化集成挑選方法在多數資料集上皆優於傳統集成挑選方法，能更有效地挑出較佳的基本模型組合，提升整體預測效能，在基本模型生成策略方面，本研究提出之隨機生成結合粒子群優化演算法，於集成挑選後在多數資料集表現優於袋裝法，而進一步結合隨機生成結合粒子群優化演算法與袋裝法之混合策略，於整體平均預測正確率上達到最佳表現，顯示整合不同模型生成機制可進一步強化分類效能。

Ensemble learning improves prediction performance by integrating the outputs of multiple base models, and ensemble selection is an effective way to reduce the computational cost in ensemble learning. Traditional ensemble selection methods consider both accuracies and diversities of base models. When base models are randomly generated, an optimization model for ensemble selection need not include the diversity among base models in its objective function. However, the optimization model is designed to perform ensemble selection for data sets with only two class values. This thesis first investigates the parameters of particle swarm optimization for generating base models with high accuracies randomly for naïve Bayesian classifier when data sets have multiple class values. Then the filtering mechanism and the optimization model for ensemble selection are redesigned for processing data sets with multiple class values. The experimental results on 20 data sets demonstrate that the ensemble selection method proposed in this study is superior to another that considers both accuracies and diversities of base models. The hybrid of randomly-generated base models with the ones induced by bagging approach achieves the highest mean classification accuracy.

黃中立，(2023)，以簡易貝氏分類器隨機生成基本模型之集成方法。國立成功大學 資訊管理研究所碩士班碩士論文。
徐心縈，(2023)，用羅吉斯迴歸建構隨機分類模型之集成方法。國立成功大學資訊管理研究所碩士班碩士論文。
蔡哲倫，(2024)，適用於多類別資料的含隨機生成模型之集成嵌套二分法。國立成功大學資訊管理研究所碩士班碩士論文。
何政賢，(2024)，以粒子群最佳化方法優化應用於二類別資料之隨機集成演算法。國立成功大學資訊管理研究所碩士班碩士論文。
陳毓潔，(2025)，適用於二類別資料之基於正確率最佳化的集成挑選法。國立成功大學工業與資訊管理研究所碩士班碩士論文
林韋成，(2025)，應用一對一與一對多方法於多類別資料的隨機生成模型之集 成方法。國立成功大學資訊管理研究所碩士班碩士論文。
Ali, M.A, Uçuncu, D., Karadayı Ataş, P., Özöğür Akyüz,S. (2019). Classification of motor imagery task by using novel ensemble pruning approach. IEEE Transactions on Fuzzy Systems, 28(1), 85-91.
Bian, Y.J., Wang, Y.J., Yao, Y.Q., Chen, H.H. (2020). Ensemble pruning based on objection maximization with a general distributed. IEEE Transactions on Neural Networks and Learning Systems, 31(9), 3766-3774.
Breiman, L. (1996). Bagging predictors. Machine Learning, 24, 123-140.
Chen D.C., Li W.W., Fang J. (2025). Blending-based ensemble learning low-voltage station area theft detection. Energies, 18(31).
Concha, B. and Pedro,L. (2014). Discrete bayesian network classifiers: A survey. ACM COMPUTING SURVEYS, 47(1), 1-43.
Dong, X.B., Yu, Z.W., Cao W.M, Shi Y.F, Ma, Q.L. (2020). A survey on ensemble learning. Frontiers of Computer Science, 14(2), 241-258.
Dong, X., Wang, J., Liang, Y. (2025). A novel ensemble classifier selection method for software defect prediction, IEEE Access, 13, 25578-25597.
Farid, D.M., Zhang, L., Rahman, C.M., Hossain, M.A., Strachan, R. (2014). Hybrid decision tree and naïve Bayes classifiers for multi-class classification tasks. Expert Systems with Applications, 41(4), 1937-1946.
Galar, M., Fernández, A., Barrenechea, E., Bustince, H., Herrera, F.(2011). An overview of ensemble methods for binary classifiers in multi-class problems: Experimental study on one-vs-one and one-vs-all schemes. Pattern Recognition, 44(8), 1761-1776.
Galar, M., Fernández, A., Barrenechea, E., Herrera, F. (2015). DRCW-OVO: Distance-based relative competence weighting combination for One-vs-One strategy in multi-class problems. Pattern Recognition, 48(1), 28-42.
Garcia, L.P.F., Sáez, J.A., Luengo, J., Lorena, A.C, Carvalho, A.C.P.L.F., Herrera, F. (2015). Using the One-vs-One decomposition to improve the performance of class noise filters via an aggregation strategy in multi-class classification problems. Knowledge-Based Systems, 91, 153-164.
Ibomoiye, D.M., Sun Y.X. (2022). A survey of ensemble learning : concepts, algorithms, applications, and prospects. IEEE Access, 10, 99129-99149.
Koutsoukas, A., Lowe, R., KalantarMotamedi, Y., Mussa, H.Y., Klaffke, W., Mitchell, J.B.O., Glen, R.C., Bender, A. (2014). In silico target predictions: Defining a benchmarking data set and comparison of performance of the multiclass naïve bayes and parzen-rosenblatt window. Journal of Chemical Information and Modeling, 54(7), 2180-2182.
Li, C.L, Li, D.Y., Qin, X., Qin, Y.X, Li, J.L., Yuan, J.F.(2025). Clustering ensemble pruning algorithm based on vec2vec classifier representation. Journal of King Saud University Computer and Information Sciences, 37(22).
Mehra, N. and Gupta S. (2013). Survey on multiclass classification methods. International Journal of Computer Science and Information Technologies, 4(4), 572-576.
Mitu, M.M., Arefin, S., Saurav, Z., Hasan M.A.,Farid, D.M. (2024). Pruning-based ensemble tree for multi-class classification. 2024 6th International Conference on Electrical Engineering and Information & Communication Technology, 481-486.
Nakata, N., Siina, T. (2023). Ensemble learning of multiple models using deep learning for multiclass classification of ultrasound images of hepatic masses. Bioengineering, 10(69).
Rokach, L. (2009). Collective-agreement-based pruning of ensembles. Computational Statistics & Data Analysis, 53(4), 1015-1026.
Sagi, O., Rokach, L. (2017). Ensemble learning : A survey. Wiley Interdisciplinary Reviews-Data Mining and Knowledge Discovery, 8(4).
Tirronen, S., Kadiri, S.R., Alku, P. (2023). Hierarchical multi-class classification of voice disorders using self-supervised models and glottal features. IEEE Open Journal of Signal Processing, 4, 80-88.
Wu, Y.C., He, Y.X., Qian, C., Zhou, Z.H. (2022). Multi-objective evolutionary ensemble pruning guided by margin distribution. Parallel Problem Solving from Nature – PPSN XVII, 427-441.
Zhang, Y., Burer, S., Street, W.N. (2006). Ensemble pruning via semi-definite programming. Journal of Machine Learning Research, 7, 1315-1338.