【摘要】 计算机辅助医学诊断是机器学习技术的一个重要实践,但是在医学诊断中一个重要影响因素来自于数据集中的冗余特征。为了消除诊断中冗余特征对集成学习方法的精度的影响,文章提出了一种PCA-FS-Bagging算法,利用主成份分析进行特征变换来解决这个问题,算法在三个医学诊断数据集上与其它算法比如单个支持向量机、支持向量机Bagging集成等进行了性能比较,结果显示了PCA-FS-Bagging算法具有较好的性能。更多还原

【Abstract】 Computer aided medical diagnosis is an important practice of machine learning techniques,but a critical factor reducing the accuracy of medical diagnosis is from the redundant features of medical data sets.In order to improve the diagnosis accuracy,PCA-FS-Bagging is proposed in this paper using principle component analysis to solve the redundant features.PCA-FS-Bagging is compared with single support vector machine,bagging of support vector machines on three real benchmark data sets.Experimental results show that PCA-FS-Bagging performs better than the other two algorithms do.更多还原