中国组织工程研究

中国组织工程研究 ›› 2010, Vol. 14 ›› Issue (52): 9803-9806.doi: 10.3969/j.issn.1673-8225.2010. 52.028

• 数字化骨科 digital orthopedics • 上一篇    下一篇

基于微量元素支持向量机在鼻咽癌模型预测中的应用

阳  春1，杜  曦2，杜  军2，唐  斌2，刘  可2，胡  昕3　　   

  1. 1中国水利水电第十工程局医院，四川省都江堰市  611830；2泸州医学院化学教研室，四川省泸州市  646000；  3泸州医学院附属第二医院，四川省泸州市  646000
  • 出版日期:2010-12-24 发布日期:2010-12-24
  • 通讯作者: 唐斌，硕士，副教授，泸州医学院化学教研室，四川省泸州市 646000 tangbin8888@163.com
  • 作者简介:阳春，男，1973年生，四川省宜宾市人，汉族，四川大学华西医学院毕业，主治医师。 76858236@qq.com
  • 基金资助:

    泸州市重点科技项目[2009-S-15(7/7)]，课题名称“基于微量元素的支持向量机在四川南部地区、贵州地区鼻咽癌模型预测中的应用”。

Support vector machine for classifying model of nasopharyngeal carcinoma based on microelement

Yang Chun1, Du Xi2, Du Jun2, Tang Bin2, Liu Ke2, Hu Xin3   

  1. 1 Hospital Affiliated to Sinohydro Bureau 10, Dujiangyan  611830, Sichuan Province, China; 2 Department of Chemistry, Luzhou Medical College, Luzhou  646000, Sichuan Province, China; 3 Second Affiliated Hospital of Luzhou Medical College, Luzhou 646000, Sichuan Province, China
  • Online:2010-12-24 Published:2010-12-24
  • Contact: Tang Bin, Master, Associate professor, Department of Chemistry, Luzhou Medical College, Luzhou 646000, Sichuan Province, China tangbin8888@163.com
  • About author:Yang Chun, Attending physician, Hospital Affiliated to Sinohydro Bureau 10, Dujiangyan 611830, Sichuan Province, China
  • Supported by:

    the Key Science and Technology Program of Luzhou, No. 2009-S-15(7/7) *

PDF

367

被引次数

4

摘要：

背景：支持向量机目前已经在文本分类、手写识别、图像分类、生物信息学等诸多领域被成功应用。
目的：采用智能算法，将支持向量机算法与微量元素数据结合对鼻咽癌患者建模，以提高鼻咽癌识别正确率。
方法：基于微量元素数据，利用支持向量机对鼻咽癌患者、正常人、其他疾病患者样本建立分类模型。样品取自观察对象未染发头枕部紧贴头皮3 cm的头发。对样本进行的临床微量元素检测项目为6种元素锌、铜、铁、锰、镉、镍，加上年龄和性别共8项。采用高斯径向基函数为核函数、调节核函数参数C及σ以建立最佳支持向量机模型。
结果与结论：采用十折交叉验证法得到模型的识别率分别为81.71%和66.47%。结果表明，基于微量元素的支持向量机法建立的鼻咽癌分类模型能较好的把鼻咽癌样本从正常人、各种疾病患者样本中区分出来。

关键词: 支持向量机, 鼻咽癌, 微量元素, 模式识别, 疾病诊断

Abstract:

BACKGROUND: Support vector machine (SVM) has been successfully used in document classification, handwriting recognition, image classification and bioinformatics.
OBJECTIVE: Using intelligent algorithm, to establish model of nasopharyngeal carcinoma (NPC) patients with SVM and microelement data to improve the identification accuracy of NPC.
METHODS: NPC was used to classified model of NPC patients and the normal or the other disease patients based on microelement data. The sample was harvested from non-colored hairs, 3 cm from the scalp. The microelement detection included zinc, copper, ferri, manganese, chromium and nickel, in addition to age and sex. The radical basis function is adopted as a kernel function of SVM, and the model adjusts C and σ to build the optimization classifier.
RESULTS AND CONCLUSION: The correct classification ratio was 81.71% and 66.47% by 10-fold cross validation. The result shows that the classified model of blood routine based on SVM can classifies the nasopharyngeal carcinoma patients from the normal or the other disease patients.

中图分类号: