Predicting Gestational Diabetes Using an Intelligent Neural Network Algorithm

Introduction: Due to the large amount of data on people with diabetes, it is very difficult to extract the predictors of diabetes. Data mining science has achieved this important goal with the help of its effective methods with the aim of discovering the prediction of diseases and has helped physicians and medical staff in predicting and diagnosing diseases.    Methods: The present research is an applied-survey type conducted in 1399. In this research, the data set of Mir Sharif et al. Has been used. Here, the primary data collection method is used to collect data and the statistical population includes 105 cases of patients registered from 1390 to 1393 in a field study of the gynecology clinic in Tehran, which of these, 80 are healthy people and 25 are people with gestational diabetes. MATLAB software has been used to analyze and evaluate the results. Results: The results and comparisons made in this study show the high efficiency of the proposed method in predicting gestational diabetes patients. Also, the accuracy of the proposed method was 93%, which was more accurate than the method of Mir Sharif et al. On the same data set. Conclusion: The proposed system has a good performance and in terms of accuracy in the data set compared to previous methods has reached 93.2%. Therefore, an intelligent and unsupervised approach can be used to diagnose gestational diabetes.