Relationship between helicobacter pylori infection and type 2 diabetes using machine learning BPNN mathematical model under community information management

Abstract

Objective: This exploration aims to explore the effect of Helicobacter pylori infection on blood glucose mechanism and gastric function in patients with type 2 diabetes under the background of electronic medicine, and the effect of community information management platform on health management efficiency of patients with Helicobacter pylori, so as to lay the foundation for clinical research on Helicobacter pylori participation in the mechanism of type 2 diabetes. Methods: In this paper, 300 patients who were treated in our hospital from June 2016 to June 2019 were selected as the research object. Among them, 228 patients with type 2 diabetes were recorded as the experimental group, and 72 patients with non-type 2 diabetes were recorded as the control group. 13C breath test and serum IgG antibody test were performed on each research object. Patients infected with Helicobacter pylori in the experimental group were divided into groups according to the degree of urine albumin excretion. Urinary albumin excretion rate was <30 mg/24 h as group A, between [30,300]mg/24 h as group B, and greater than 300 mg/24 h as group C. The control group was recorded as group D. Blood biochemical indexes and gastroscopy were detected in the four groups; the blood biochemical indexes of each group were compared and analyzed by statistical software; the artificial intelligence health information platform in community information management was established, and the mathematical prediction model of diabetes was established based on Back Propagation Neural Network (BPNN). Results: The proportion of Helicobacter pylori infection in patients with type 2 diabetes was 60%, and the proportion of Helicobacter pylori infection in the control group was 40%. There was a significant difference in fasting blood glucose indicator and cholesterol indicator between group C and group D, P < 0.05. There was a significant difference in the percentage indicator of glycated protein between group A and group C, P < 0.05. There was a significant difference in normal gastroscopy between group A and group D, P < 0.05. In the process of training, the error of the train set of the mathematical model based on BPNN is gradually reduced, and it has good convergence. When the number of hidden layer units is 3, the AUC (Area Under Curve) of train set is the largest. When the number of hidden layer units is 1, the AUC of the test set is the largest, so the network model with one hidden layer unit is selected. Conclusion: The community use efficiency of each performance of the artificial intelligence health information platform in community information management has been significantly improved compared with that before optimization, which can improve the health management level on the basis of patients' electronic medical information. © 2021 The Author(s)

Description

Keywords

Artificial intelligence, BPNN, Community information management health information platform, Helicobacter pylori, Mathematical model, Type 2 diabetes

Citation

Ma, H., Xiao, J., Chen, Z., Tang, D., Gao, Y., Zhan, S., Ghonaem, E., & Abo Keir, M. Y. (2021). Relationship between Helicobacter Pylori Infection and Type 2 Diabetes Using Machine Learning BPNN Mathematical Model under Community Information Management. Results in Physics, 26, 104363. https://doi.org/10.1016/j.rinp.2021.104363

DOI