基于标准数据集的分类器融合学习模型

Classifier Fusion Learning Model Based on Standard Dataset

判别式分类器通过生成不同复杂度的指示函数去调节算法与所解决问题的适应性,能有效地避免过拟合现象。分类器融合方法就是应用单个分类器对特定样本预报的特异性来提高模型的整体预测精度,应用支持向量机(SVM)对乳腺癌数据进行建模,通过选取不同的模型参数(径向基核函数参数gamma和正则化约束参数cost)构建9个单分类器,通过投票策略在单分类器上构建融合分类器,融合模型对乳腺癌数据的预测精度为98.59%,相比单分类模型对此数据集的预测精度97.72%有明显的竞争力,试验结果表明融合模型能有效提升分类器的泛化能力。

The discriminant classifier generates indicators with different complexitres that adjusts flexibility between method and problems, which can efficiently avoid the over-learning. Fusion method is to improve the prediction accuracy by summarizing the specificities of individual classifiers. The purpose of the study is to predict breast cancer with support vector machine(SVM). Nine individual classifiers are trained by selecting different parameters(gamma of radial basis function, cost of regularization parameter), on which the fusion classifier is construct by using voting strategy. 98.59% prediction accuracy is obtained, it is very promising compared with 97.72% obtained by optimal individual classifier. The experimental results indicate that the ensemble model can enhance the prediction accuracy.