The diagnostic model of diabetic retinopathy was established
小组成员：郭美婷、何佳倩、王松、王笙璎、阿克那尔
指导老师：何俊
糖尿病是一种以高血糖为特征的代谢性疾病，长期高血糖会慢性损害各种器官。糖尿病视网膜病变(DR)便是糖尿病常见的并发症之一。若不进行干预任其发展，患者最后的结局就是不可逆盲。大多数患者在出现黄斑水肿或增殖性糖尿病视网膜病变之前没有症状，因此，及时筛查糖尿病视网膜病变可以保护糖尿病患者视力。我们拟使用机器学习方法构建糖尿病视网膜病变诊斯模型，可及时筛查糖尿病视网膜病变。
图片源于知乎@眼科童医生
Training Set
(Data from GSE221521,blood samples were collected in 50 healthy controls, 74 diabetic patients without diabetic retinopathy (DM group), and 69 diabetic retinopathy (DR Group).After feature selection, the following feature columns are obtained.)
Test Set
(Data from GSE221521,blood samples were collected in 50 healthy controls, 74 diabetic patients without diabetic retinopathy (DM group), and 69 diabetic retinopathy (DR Group).After feature selection, the following feature columns are obtained.)
downloadModel evaluation data
Below are the evaluation data and corresponding histogram visualizations of the five models we have tried, namely random forest, vector machine, decision tree, Bayesian and logistic regression models, etc. According to the evaluation data, we chose the vector machine model for subsequent experiments.
