基于多元线性回归的糖尿病治疗效果评价模型

Evaluation Model of Diabetes Therapeutic Effect Based on Multiple Linear Regression

摘要: 本文针对糖尿病患者的再次入院率与其相关特征向量的关系进行研究，采用岭回归以及线性拟合得到评价指标体系，通过逐步回归得到了再次入院率与其相应特征变量之间的关系模型。对模型结果分析表明：医院对糖尿病患者治疗成功率为74.32%；中低水平收入者可考虑选用insulin、metformin对糖尿病进行控制；高水平收入者可考虑选用tolazamide对糖尿病进行控制。本问题出的治疗模型有助于糖尿病患者的合理治疗。

Abstract: This paper studied the relationship between the readmission rate of diabetic patients and its related eigenvectors. The diagnostic model was made by ridge regression and linear fitting. The relationship between the readmission rate and its corresponding characteristic variables was obtained by stepwise regression. The analysis of model results shows that: the success rate of diabetes treatment in hospitals is 74.32%; insulin and metformin can be used to control diabetes in low and middle income groups; tolazamide can be used to control diabetes in high income groups. The treatment model of this problem is helpful to the rational treatment of diabetes mellitus patients.