The Research on Factors Influencing Diabetes

Abstract

Diabetes is a serious disease for people, causing massive damage to body. However, people learn a little about the diabetes, showing that it is very important to analyze the relationships between diabetes and several factors. We incorporate multiple factors such as age and gender into a unified framework to jointly test the impact of multiple factors on the dependent variable. Whether a person owning the diabetes is been seen as the result. This paper established logistic regression model and using a comprehensive dataset sourced from Kaggle containing these factors to estimate paraments and using F1-score and Receiver Operating Characteristic (ROC) curve to evaluate the whole model. The results show that the estimation of parameter HbA1c is 2.53, playing the most important role in the diabetes. Blood glucose, Age, and BMI account, either. Apart from that, it could be learned from the Stand Error (SE) and P-value about the accuracy. The characters of F1-score and ROC curve show that this model performs exceptionally well in identifying diabetes cases. A conclusion could be drawn that a person could try to exercise or improve diet to release the symptom of diabetes.