Author + information
Sex hormone levels in elderly male patients with type 2 diabetes mellitus (T2DM) and hypertension were analyzed based on a classical k-means clustering algorithm to explore the correlation between serum sex hormone levels, blood lipid (including triglyceride, cholesterol, low density lipoprotein, high density lipoprotein, apolipoprotein A and apolipoprotein B), insulin, and blood tension.
600 cases of clinical data were randomly selected from the hospital's existed database including 400 cases as the test group and 200 cases as the control group. A high dimension clinical data model such as height, weight, age, waist circumference, pulse rate, heart rate, systolic blood pressure, diastolic blood pressure, triglyceride (TC), low density lipoprotein (LDL), high density lipoprotein (HDL), fasting blood glucose, total cholesterol (CH), serum androgen (T), estrogen (E2), and insulin was established. The data were processed by statistical methods, and the initial clustering center points were determined according to the characteristic density. In addition, the initial center points should be evenly distributed. The different norm was selected to measure the distance between the cluster initial center points and the different data objects. The data objects were assigned to the class where each cluster center point was located according to the minimum principle of distance. Then the mean of each class was calculated as the next iteration of the clustering center point.
The list of the rapid clustering analysis showed that the data set was divided into three classes. The E2 level in elderly males with T2DM was higher than that in normal group while T is declining, and the ratio of E2/T was significantly higher than that of the control group. Serum E2 level in patients with diabetes mellitus complicated with hypertension was significantly lower than that in normal diabetic group, E2/T is still higher than the normal blood pressure group; hypertensive patients with insulin, total cholesterol was significantly higher than normal blood pressure group. The center table showed that the final center positions of the center points has changed greatly from the initial center positions. The differences between the indices (mean variance and degree of freedom) of each data set were obvious. The iteration history table shows that the iterative process of this clustering is fast.
Elderly male patients with type 2 diabetes mellitus suffered from significant sex hormone abnormalities, lipid metabolism disorders and insulin resistance. It may be involved in the formation of diabetes with hypertension.