摘要
通过分析检查者各属性情况,研究对其患糖尿病概率的影响并制定相应的风险指标。利用检查者数据作为研究数据,用各属性的相关系数值来分析选取可能影响检查者存活概率的属性,并对所选取的属性分别进行单个属性的一元logistic回归和多个属性的多元logistic回归,并加入标准化后的HbA1c值联合患病概率考虑风险系数。多元logistic回归分析显示,性别、年龄γ-谷氨酰基转移酶,血清白蛋白,钙属性较大影响检查者患病概率的因素。认为检查者的患病概率并不是随机的,而是受检查者多个属性所影响的。
To study the influence on the probability of diabetes mellitus (DM) of examinees by analyzing the attributes of examinees, using the data of examiners as the research data, the correlation coefficient value of each attribute was used to analyze and select the attributes that may affect the survival probability of examiners, and the single logistic regression of single attribute and multiple logistic regression of multiple attributes were carried out for the selected attributes, and the standardized HbA1c value was added to consider the risk coefficient. Multiple logistic regression analysis shows that sex, age, y-glutamyltransferase, serum albumin, calcium attribute have a great influence on the probability of disease. It is not random, but influenced by multiple attributes of examinees.
出处
《运筹与模糊学》
2020年第4期321-329,共9页
Operations Research and Fuzziology
