| 引用本文:王娟1，赵慧辉1，陈建新1，罗良涛2，李雪丽1，王金平1，刘俊杰1，王伟1.基于神经网络技术分析的慢性心力衰竭血瘀证诊断模型研究[J].世界中医药,2018,(09):. |
|
| 基于神经网络技术分析的慢性心力衰竭血瘀证诊断模型研究 |
| Study on Diagnosis Model of Chronic Heart Failure Patients with Blood Stasis Syndrome Based on Neural Network Analysis |
| 投稿时间：2018-06-29 |
| DOI：10.3969/j.issn.1673-7202.2018.09.006 |
| 中文关键词: 慢性心力衰竭 血瘀证 诊断模式 数据挖掘 |
| English Keywords:Chronic heart failure Blood stasis syndrome Diagnostic mode Neural network analysis |
| 基金项目:国家重点研发计划资助，气虚证辨证标准的系统研究（2017YFC1700100）；国家自然基金项目资助（81302914）——“基于宏观表征与微观指标组合的中医证候诊断方法研究” |
|
| 摘要点击次数: 931 |
| 全文下载次数: 666 |
| 中文摘要: |
| 目的：应用多元数据统计和神经网络挖掘方法构建慢性心力衰竭（Chronic Heart Failure，CHF）患者血瘀证基于多系统理化指标信息的诊断模型，探索与慢性心力衰竭患者血瘀证相关的理化指标信息的组合模式及其生物学意义。方法：选取2010年3月至2011年8月四川省成都市中西医结合医院和浙江省杭州市中医院收治的CHF患者100例，分为观察组（血瘀证患者组）和对照组（非血瘀证患者组），其中观察组37例，对照组57例。观察组为进行临床流行病学调查，收集患者四诊信息和生物样本进行多系统理化指标的检测分析。在分析差异指标基础上，综合应用回归方法及神经网络数据挖掘方法进行数据模型建设，从而形成慢性心力衰竭血瘀证患者的多系统理化指标诊断模型。结果：本研究共纳入100例心力衰竭患者，其中血瘀证患者占37%。应用上述方法筛选出有统计学意义的多系统理化指标信息20项，按照条目的显著性顺序依次进入神经网络数据挖掘模型，显示筛选指标的建模准确率为75.4%，测试样本的准确率达到82.4%。结论：神经网络数据挖掘方法可以用于临床理化指标信息数据进行CHF患者血瘀证的建模分析，与血瘀者密切相关的多系统理化指标信息也为进一步了解心力衰竭血瘀证的病理生理机制提供了参考依据。 |
| English Summary: |
| To construct the diagnosis model of chronic heart failure (CHF) patients with blood stasis syndrome based on multi-system physical and chemical index information by multiple data statistics tools and neural network mining methods,and to explore the biological significance of the physical and chemical index information related to the blood stasis syndrome of chronic heart failure patients.Methods:Clinical epidemiological investigation of 100 patients with CHF treated in Chengdu Integrated TCM & Western Medical Hospital and Hangzhou Hospital of Traditional Chinese Medicine from March 2010 to August 2011 was conducted,and the four diagnosis information of the patients was collected.The physical and chemical indexes of the biological samples were also detected and analyzed.On the basis of analysis of difference index,the regression method and the neural network data mining method were used to analyze the data to form the multi-system physical and chemical index diagnosis model of the patients with chronic heart failure and blood stasis syndrome.Results:The study included 100 patients with heart failure,of which 37% were patients with blood stasis syndrome.The above method was used to screen out 20 items of multi-system physical and chemical index,and the data mining model of the neural network was conducted according to the order of significance of the items.The accuracy of the model was 75.4%,and the accuracy of the test sample was 82.4%.Conclusion:The neural network data mining method can be used in the modeling and analysis of CHF patients with the blood stasis syndrome based on clinical physical and chemical information data.The multi-system physical and chemical information closely related to blood stasis also provide reference for further understanding of pathophysiological mechanism of chronic heart failure and blood stasis syndrome. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
