基于因子分析的二元Logistic回归对糖尿病预测的研究
Prediction of Diabetes by Binary Logistic Regression Based on Factor Analysis

目前全球糖尿病患者数量庞大，且数据在持续增长，有近一半人不知晓自己已经患病。为了尽早诊断及时治疗防控，本文建立了一个糖尿病初步诊断模型。本文选取UCI机器学习库中的糖尿病数据集，利用SPSS软件进行数据分析。首先利用因子分析方法对影响因素数据进行降维处理，得到公共因子和其对应数值，然后根据公共因子建立二元Logistic回归模型，根据模型可以预测是否患病。
At present the number of diabetics worldwide is huge and growing, and nearly half of people don’t know that they are sick. In order to diagnose and treat diabetes as soon as possible, this article establishes a preliminary diagnosis model of diabetes. We select diabetes data set from UCI machine learning library and use SPSS software for data analysis. First, the factor analysis method is used to reduce the dimension of the influencing factor data, and the common factors and their corresponding values are obtained. Then, a binary Logistic regression model is established according to the common factors, and the disease can be predicted according to the model.