基于支持向量机的急性出血性脑卒中早期预后模型的建立与评价

发布时间：2018-02-06 04:31

本文关键词： 急性出血性脑卒中预后支持向量机 Logistic回归　出处：《南京医科大学学报(自然科学版)》2016年01期 　论文类型：期刊论文

【摘要】：目的:比较支持向量机(support vector machine,SVM)和传统的Logistic回归构建的急性出血性脑卒中(intracerebral hemorrhage,ICH)早期预后判别模型的预测性能,探索急性ICH预后研究的新方法。方法 :收集急性ICH患者339例,随访观察21 d时的临床转归情况。应用随机数字法以3∶1的比例分为两组,一组作为训练样本用于筛选变量和建立预测模型,计254例;另一组作为验证样本,用于评价模型预测效果,计85例。建模方法采用SVM和常规统计方法中的Logistic回归。结果:通过对85例ICH患者的预测判别验证,SVM1的预测分类能力在4个模型为最强,4个模型预测的准确率和Youden指数分别为:Logistic回归:72.9%(62.0%~81.7%)、0.441(0.249~0.633);SVM1:82.4%(72.3%~89.5%)、0.632(0.465~0.799);SVM2:78.8%(68.4%~86.6%)、0.557(0.379~0.735);SVM3:78.8%(68.4%~86.6%)、0.563(0.385~0.741)。结论:采用SVM能较好地判断急性ICH患者的早期预后,其效能优于Logistic回归模型。
[Abstract]:Objective: to compare support vector machine with support vector machine. SVM) and traditional Logistic regression were used to construct acute hemorrhagic stroke / intracerebral hemorrhage. Methods: 339 patients with acute ICH were collected. The clinical outcome of 21 days was followed up and divided into two groups with random number method of 3: 1. The first group was used as training sample to screen variables and establish prediction model for 254 cases. The other group is used as validation sample to evaluate the prediction effect of the model. 85 cases were included. SVM and Logistic regression were used in the modeling method. Results: 85 cases of ICH patients were predicted and verified. The prediction and classification ability of SVM1 is the strongest in the four models. The prediction accuracy and Youden index of the four models were respectively: 72.9% and 62.0% and 81.7%, respectively. 0.441U 0.249C 0.633m; SVM1: 82.4 and 72.3%, 89.5and 0.6320.4650.4650.99m; SVM2: 78.8 and 68.4, including 0.5570.379and 0.735; SVM 3: 78.8% and 68.4%. Conclusion: the early prognosis of patients with acute ICH can be judged better by using SVM. Its efficiency is superior to that of Logistic regression model.
【作者单位】：南京医科大学公共卫生学院;南京中医药大学中医统计研究和咨询中心;南京中医药大学第一临床医学院;
【基金】：国家自然科学基金资助项目(81373512) 国家重点基础研究发展计划(973)资助项目(2006CB504807) 江苏高校优势学科资助项目(PAPD)
【分类号】：R743.3;R-332

【相似文献】