研究了基于聚类技术提高分类器差异性的方法.通过Bootstrap技术与分类器学习算法训练分类器模型,利用分类器在验证集上的分类结果作为聚类的数据对象;然后应用聚类算法对这些数据聚类,并在每个簇中选择分类器代表模型,以此构成集成学习的成员;最后应用融合方法实验研究了基于聚类技术提高差异性的集成学习性能,并与集成学习方法bagging,adaboost进行了实验比较.
[1] OPITZ D, SHAVLIK J. Generating accurate and diverse members of a neural-network ensemble [M]. Cambridge, MA:The MIT Press 1996.
[2] ROSEN B E. Ensemble learning using decorrelated neural networks [J]. Connection Science 1996, 8(03).
[3] Zhi-Hua Zhou, Jianxin Wu, Wei Tang. Ensembling neural networks: Many could be better than all [J]. Artificial Intelligence: An International Journal 2002, 1/2(1/2).
[4] Altincay H. Decision trees using model ensemble-based nodes [J]. Pattern Recognition: The Journal of the Pattern Recognition Society 2007, 12(12).
[5] Ko AHR, Sabourin R, Britto AD, Oliveira L. Pairwise fusion matrix for combining classifiers [J]. Pattern Recognition: The Journal of the Pattern Recognition Society 2007, 8(8).
[6] SIRLANTZIS K, HOQUE S, FAIRHURST M C. Diversity in multiple classifier ensembles based on binary feature quantisation with application to face recognition [J]. Applied Soft Computing Journal 2008, 8(01).
[7] GIACINTO G, ROLI F. Design of effective neural network ensembles for image classification purposes [J]. Image and Vision Computing 2001, 19(9-10).
[8] Miin-Shen Yang, Kuo-Lung Wu. A similarity-based robust clustering method [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence 2004, 4(4).
[9] YU Jian, HUANG Houkuan. A new weighting fuzzy C-means algorithms [A]. St Louis Missouri 2003.
[10] BLAKE C L, MERZ C J. UCI Repository of Machine Learning Database [EB/OL]. http://www.ics.uci.edu/～mlean/MLRepository.html 2008.
