%0 Journal Article %T 基于机器学习分析NHR与冠心病的相关性
Correlation between Neutrophil-to-High-Density Lipoprotein Cholesterol Ratio (NHR) and Coronary Heart Disease Based on Machine Learning Analysis %A 郭萌萌 %A 蔡叶锐 %A 姜泽军 %A 杨红玲 %J Advances in Clinical Medicine %P 150-162 %@ 2161-8720 %D 2025 %I Hans Publishing %R 10.12677/acm.2025.152328 %X 目的：探讨中性粒细胞与高密度脂蛋白胆固醇比值(NHR)对冠心病(CHD)的预测价值。方法：研究940名研究对象，包括548例冠心病组以及392例对照组。相关性分析计算CHD与NHR的相关系数。最小绝对收缩和选择算子(LASSO)回归分析筛选潜在影响。构建机器学习模型进行评估变量在诊断的重要性。采用ROC曲线分析危险因素对冠心病的预测能力。使用Spearman相关性分析来评估NHR与Gensini评分之间的关系。结果：相关性分析显示冠心病与NHR的相关系数为0.25，呈正相关(P < 0.001)。LASSO回归分析显示，NHR与冠心病风险显著相关。使用机器学习算法进行进一步分析发现，NHR对冠心病的诊断具有良好的预测价值。ROC曲线分析显示，特征重要性排名前7个的危险因素的组合分析提高了CHD的预测价值(AUC值：0.738，P < 0.001)。此外，Spearman相关性分析发现随着NHR水平的增加，Gensini评分逐渐增加(r = 0.32, P < 0.001)。结论：NHR是发生CHD的独立危险因素，与CHD的预测和严重程度密切相关。XGBoost模型对CHD与NHR具有良好的应用价值。
Objective: To investigate the predictive value of the neutrophil-to-high-density lipoprotein cholesterol ratio (NHR) for coronary heart disease (CHD). Methods: A total of 940 subjects were studied, including 548 in the CHD group and 392 in the control group. Correlation analysis was conducted to calculate the correlation coefficient between CHD and NHR. LASSO (Least Absolute Shrinkage and Selection Operator) regression analysis was performed to identify potential influencing factors. A machine learning model was constructed to evaluate the importance of variables in diagnosis. ROC (Receiver Operating Characteristic) curve analysis was used to assess the predictive ability of risk factors for CHD. Spearman correlation analysis was employed to evaluate the relationship between NHR and Gensini score. Results: Correlation analysis showed a positive correlation between CHD and NHR, with a correlation coefficient of 0.25 (P < 0.001). LASSO regression analysis revealed a significant association between NHR and CHD risk. Further analysis using machine learning algorithms demonstrated that NHR has good predictive value for CHD diagnosis. ROC curve analysis indicated that the combination of the top 7 most important risk factors improved the predictive value for CHD (AUC: 0.738, P < 0.001). Additionally, Spearman correlation analysis found that as NHR levels increased, the Gensini score also increased (r = 0.32, P < 0.001). Conclusion: NHR is an independent risk factor for CHD and is closely associated with both the prediction and SEVERITY of CHD. The XGBoost model shows good applicability for CHD and NHR. %K 冠心病， %K 中性粒细胞与高密度脂蛋白胆固醇比值， %K 机器学习
Coronary Heart Disease %K Neutrophil-to-High-Density Lipoprotein Cholesterol Ratio %K Machine Learning %U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=106839