一种用于乳腺癌诊断的免疫分类算法

来源期刊:中南大学学报(自然科学版)2010年第4期

论文作者:邓泽林 谭冠政 叶吉祥 范必双

文章页码:1485 - 1490

关键词:人工免疫识别系统;核函数;分类算法;医疗诊断;乳腺癌

Key words:artificial immune recognition system; kernel function; classification algorithm; medical diagnosis; breast cancer

摘    要:基于人工免疫识别系统AIRS(Artificial immune recognition system)和核函数提出免疫分类算法Kernel-AIRS。Kernel-AIRS遵循AIRS算法框架,利用核函数将输入空间投影到高维核空间,以核空间距离来度量抗体-抗原的亲和度,提高算法对非线性可分问题的分类准确率。采用Kernel-AIRS定义核空间距离测量方法和规一化方法,分析抗体刺激度和核函数参数与分类准确率之间的关系,研究属性缺失样本对算法分类准确率的影响,并应用Kernel-AIRS算法诊断乳腺癌,分类准确率采用10次交叉验证评价。研究结果表明:Kernel-AIRS算法对排除属性缺失样本数据集分类的准确率为97.3%,对包含属性缺失样本数据集分类的准确率为96.9%,分类准确率较高,适用于乳腺癌的诊断。

Abstract: Based on an artificial immune recognition system (AIRS) and a kernel function, Kernel-AIRS algorithm was proposed, which followed the AIRS algorithm framework. Using Kernel function to project input space into high dimensional kernel space, the affinity of antibody and antigen was calculated in the kernel space to improve the classification accuracy for the non-linear problems. Kernel space distance computation method and normalization method were defined by Kernel-AIRS, the relationship between the parameters of simulation level and kernel function and classification accuracy was analyzed, and the effect on the classification accuracy of missing attribute samples was discussed. Kernel-AIRS was applied to diagnosis breast caner samples in Wisconsin breast cancer dataset (WBCD), and the classification accuracy was assessed by 10 fold cross validation. The results show that, using this algorithm, the accuracy rates are 97.3% and 96.9% for excluding missing attributes samples and including missing attributes samples respectively, and the classification accuracy is high and applicable for breast cancer diagnosis.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号