摘要
采用支持向量机、K-近邻法(K-NearestNeighbor,K-NN)、概率神经网络(ProbabilisticNeuralNetwork,PNN),结合乳腺肿瘤的细针穿刺细胞病理学临床数据诊断乳腺癌。结果表明:当使用Sigmoid核函数时,SVM通过5次交叉验证的最佳平均分类准确率达到了96.24%,优于K-NN(95.37%),PNN(95.09%)等分类器,表明该方法有望成为一种实用的乳腺癌临床辅助诊断工具。
Combined with the breast fine needle aspiration cytology, the SVM, K-Nearest Neighbor (K-NN) and Probabilistic Neural Network (PNN) are used to diagnose the breast cancer. The best overall accuracy reaches 96.24% via SVM with Sigmoid kernel by using 5-fold cross validation, and is superior to those of other classifiers including K-NN (95.37%) and PNN (95.09%). Support vector machine is capable of being used as a potential application tool for SVM-aided clinical breast cancer diagnosis.
出处
《重庆大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2007年第6期140-144,共5页
Journal of Chongqing University
基金
重庆市自然科学基金资助项目(CSTC
2006BB5240)
重庆大学与新加坡国立大学国际联合科研资助项目(ARF-151-000-014-112)
关键词
支持向量机
K-近邻法
概率神经网络
乳腺癌
诊断
模式识别
support vector machine
K-nearest neighbor
probabilistic neural network
breast cancer
diagnosis
pattern recognition