A Decision Support System (DSS) for Symptom-Based Differential Diagnosis in Diabetes

Diabetes, characterized by elevated blood sugar levels due to insulin deficiency or resistance, affects individuals across all age groups and necessitates lifelong management. Type 1 and type 2 diabetes present distinct symptoms, with type 1 often appearing suddenly and type 2 potentially remaining undetected. While traditional diagnostic methods are commonly utilized, limited access to healthcare can impede their efficacy, highlighting the need for innovative solutions. To solve this challenge, a predictive and user-friendly application providing preliminary diabetes information was developed using various machine learning algorithms. Among these, logistic regression emerged as the primary classification model, achieving the highest accuracy score of 0.94. These findings underscore the reliability and potential practical significance of the application in healthcare settings.