基于随机森林算法的高尿酸血症危险因素分析

目的采用随机森林算法探讨凉山地区人群高尿酸血症(HUA)患病危险因素。方法利用中国达能膳食营养研究与宣教基金项目(DIC2013-03)数据,用随机森林模型对单因素分析有统计学意义的自变量进行重要性排序并降维,将袋外估算误差率最小的变量集纳入logistic回归模型,分析自变量的作用方向和相对危险度。结果逐步随机森林分析显示,变量数为6时袋外估算误差率最低,重要性排名前六的变量依次是年龄、体质指数(BMI)、每日菌菇类摄入量、性别、每日禽畜肉摄入量、高甘油三脂(TG)血症;logistic回归分析显示,18~34岁、≥65岁组人群HUA患病风险分别是35~44岁组的1.557、1.496倍;男性HUA患病风险是女性的2.755倍;超重(24≤BMI<28)、肥胖组(BMI≥28)HUA患病风险分别是正常BMI组的1.822、2.534倍;高TG血症组HUA患病风险是非高TG血症组的2.379倍;菌菇类高摄入组(大于10g/d)HUA患病风险是低摄入组(小于5g/d)的1.420倍;畜禽肉高摄入组(大于75g/d)HUA患病风险是低摄入组(小于40g/d)组的1.300倍。结论影响凉山地区人群HUA患病的前六位危险因素依次是年龄、BMI、每日菌菇类摄入量、性别、每日禽畜肉摄入量、高TG血症,建议加大健康生活方式、饮食习惯的宣教力度,加强对BMI、高TG血症的检测和控制。

Risk Factors for Hyperuricemia Based on Random Forest Algorithm

Objective To explore the risk factors of hyperuricemia(HUA)in the population of Liangshan based on random forest algorithm.Methods Using the data of China Danone Dietary Nutrition Research and Education Fund Project(DIC2013-03),the random forest model was used to sort the importance and reduce dimension of the independent variables that were statistically significant in univariate analysis,the variable set with the smallest error rate was included in the logistic regression model to analyze the action direction and relative risk of independent variables.Results The results of the stepwise random forest analysis showed that the estimated error rate was the lowest when the number of variables was 6,the top six variables in the important ranking chart were age,BMI,daily mushroom intake,sex,daily livestock and poultry intake,high TG.Logistic regression analysis showed that the risk of HUA in the 18~34 years old and≥65 years old group was 1.557 and 1.496 times that of the 35~44 years old group,respectively.The risk of HUA in male was 2.755 times that of women.The risk of HUA in the overweight(24≤BMI< 28)and obese group(BMI≥28)was 1.822 and 2.534 times that of the normal BMI group,respectively.The risk of HUA in the high TG group was 2.379 times that of the non-high TG group.The risk of HUA in the high intake group of mushroom(greater than 10 g/d)was 1.420 times that of the low intake group(less than 5 g/d).The risk of HUA in the high intake group of livestock and poultry(greater than 75 g/d)was 1.300 times that of the low intake group(less than 40 g/d).Conclusion The top six risk factors of HUA in Liangshan area were age,BMI,daily mushroom intake,sex,daily livestock and poultry intake,high TG.It is recommended to increase the education of healthy lifestyles and eating habits,and strengthen the detection and control of BMI and high TG.