Abstract

Aim: It is a known fact that diabetes mellitus is increasing frequently and triggering many different diseases. Therefore, early diagnosis of the disease is important. This study was trying to predict the early diagnosis of the disease, according to machine learning methods by measuring plasma glucose concentration, serum insulin resistance, and diastolic blood pressure.Material and Methods: In the study, the public dataset from a website consists of 768 samples and nine variables. Three different machine learning strategies were used in the early diagnosis of diabetes mellitus (Support Vector Machine, Multilayer Perceptron, and Stochastic Gradient Boosting). 3 repeats and 10 fold cross-validation method was used to optimize the hyperparameters. The model’s performance parameters were evaluated based on accuracy, specificity, sensitivity, confusion matrix, positive predictive value (precision), negative predictive value, and AUC (area under the ROC curve).Results: According to the experimental results (the criteria of accuracy (0.79), sensitivity (0.57), specificity (0.91), positive predictive value (0.79), negative predictive value (0.80), and AUC (0.74)) the Support Vector Machine was more successful than other methods.Conclusion: Plasma glucose concentration, serum insulin resistance, and diastolic blood pressure markers are important indicators in the early diagnosis of diabetes mellitus. In this study, it was seen that these markers make a significant contribution to the early diagnosis of diabetes mellitus. However, it has been observed that these indicators alone will not be sufficient in the early diagnosis of the disease, especially since age, body mass index and pregnancy contribute significantly. 

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call