论文部分内容阅读
目的建立基于Bayes分类器的肺癌预测模型,探讨并评价该模型的预测效果。方法以前期筛选出的6个噬菌体展示肽与90例肺癌患者血清及90例正常对照血清的反应数据为基础,应用BinReg 2.0软件实现数据分析,建立Bayes肺癌预测模型,并利用受试者工作特征曲线(ROC曲线)评价比较Bayes预测模型与Logistic回归模型、主成分回归模型、支持向量机模型的分类预测效果。结果 Bayes肺癌预测模型的灵敏度为92.00%,特异度为96.00%,能够较好地区分肺癌患者与正常对照。结论 Bayes数学预测模型可较准确地预测受检者患肺癌的概率。
Objective To establish a prediction model of lung cancer based on Bayesian classifier, and to explore and evaluate the prediction effect of this model. METHODS: Based on the response data of six previously selected phage display peptides and 90 serum samples of lung cancer patients and 90 normal control serums, BinReg 2.0 software was used for data analysis to establish a Bayesian lung cancer prediction model and to use the subject’s working characteristics. Curve (ROC curve) evaluation compares Bayesian prediction model with Logistic regression model, principal component regression model, and support vector machine model. Results The Bayes’ lung cancer prediction model had a sensitivity of 92.00% and a specificity of 96.00%. It can distinguish lung cancer patients from normal controls. Conclusion The Bayes mathematical prediction model can accurately predict the probability of lung cancer in the subject.