【摘要】 目的利用支持向量机-微量元素法建立胃癌的模式识别,以期作为胃癌诊断的辅助手段。方法采用电感耦合等离子体发射光谱法与支持向量机分类模式,判别健康者与患者血清中Zn、Fe、Cu、Ni、Co、Mo、Mn、Cr和Zn/Cu 9个指标,建立胃癌诊断的模型。结果多项式核函数和径向基核函数的平均总正确率分别为91.11%和93.33%,灵敏度和特异度也达到了良好的水平。结论本方法适于辅助诊断患者胃癌疾病,并为进一步研究微量元素与各种疾病之间的关系奠定了基础。更多还原

【Abstract】 Objective To establish the identification model of gastric cancer by the support vector machine-trace element method as the supplementary means of the diagnosis of gastric cancer.Methods The inductively coupled plasma atomic emission spectrometry and support vector machines for pattern classifi cation were adopted to make discrimination for 9 indicators of serum Zn、Fe、Cu、Ni、Co、Mo、Cr and Zn/Cu between healthy volunteers and patients,and establish the diagnosis model of gastric cancer.Results The average total accuracy of polynomial kernel function and RBF kernel function were 91.11% and 93.33%,respectively.The sensitivity and specifi city also reached a good level.Conclusion This method is suitable for aided diagnosis of gastric cancer,and also lays a foundation for the further study on the relationship between trace elements and various diseases.更多还原