Sai Prakash S, A. C. Subhajini
Diabetes mellitus (DM) is aunceasingillness that causesimbalances in glucose level in blood due to the body's reluctance in generating insulin hormone. Because of its high morbidity, it has become a growing worry, and the average age of individuals affected by this disease has now dropped to the mid-twenties. Given its prevalence, it is critical to address this issue effectively. Many academics and doctors have now developed AI-based detection approaches to better tackle problems that are ignored owing to human errors. ML and DL approaches have been utilised to predict diabetes and its consequences in recent years. This research provides a DL strategy for diagnosing DM using CNN-Bi-LSTM.The approach entails retrieving essential elements from a dataset of diabetes clinical records and feeding them into a DNN. The network is then trained to recognise diabetes-related patterns in the data. The model is tested using a distinct dataset. The tests are carried out using the PIDD dataset, which contains 768 record and 8 critical variables connectedwith diabetes, each with a grouptag indicating the result of non-diabetic and diabetic individuals. The primary goal of this study is to maximise the model's accuracy.
Diabetes, Machine learning, deep learning, PIDD dataset, CNN-Bi-LSTM