Abstract
目的:利用决策树方法探讨和分析行为习惯和常见疾病协同影响颈动脉硬化的模式。方法:选取2019年7月至2020年7月于青岛大学附属医院查体并符合入组标准的人员573例,其中颈动脉硬化组329例,非颈动脉硬化组244例。收集人口学因素、行为习惯因素、伴发疾病因素、颈动脉超声指标等。采用卡方检验、t检验、logistic回归、决策树计算等方法,分析行为习惯和常见疾病协同影响颈动脉硬化的模式。结果:卡方检验及T检验显示年龄、吸烟、饮酒、高脂饮食、体育锻炼、高脂血症、高血压、糖尿病与颈动脉硬化的发生相关;logistic回归显示年龄、高脂饮食、高脂血症、体育锻炼、高血压与颈动脉硬化呈线性相关关系;决策树模型显示高血压是根节点影响因素,年龄、吸烟、体育锻炼、高脂饮食、高脂血症、糖尿病、饮酒等因素对颈动脉硬化的发生有不同的协同影响方式。ROC曲线下面积对比证实决策树模型优于logistic回归分析(p Purpose: Decision tree method was used to explore and analyze the cooperative influence of behavior habits and common diseases on carotid atherosclerosis. Methods: From July 2019 to July 2020, 573 participants (329 CA and 244 non-CA) were recruited from the Affiliated Hospital of Qingdao University. Demographic information, behavioral habits, concomitant disease factors and carotid ultrasound results were collected. Chi-square test, t-test, logistic regression and decision tree algorithm were used to analyze the cooperative influence of behavioral habits and common diseases on carotid atherosclerosis. Results: Chi-square test and T-test showed that age, smoking, alcohol consumption, high-fat diet, physical exercise, hyperlipidemia, hypertension and diabetes were correlated with the occurrence of carotid atherosclerosis. Logistic regression showed a linear correlation between age, high-fat diet, hyperlipidemia, physical exercise, hypertension and carotid atherosclerosis. The decision tree model showed that hypertension was an influencing factor at the root node, and age, smoking, physical exercise, high-fat diet, hyperlipidemia, diabetes, alcohol consumption and other factors had different synergistic influences on the occurrence of carotid atherosclerosis. The ROC curve was drawn to evaluate the predictive ability of the decision tree model (p