Gated Recurrent Unit for Diabetes Prediction in Voice Data

The burden of diabetes is increasing in India day to day. According to Civicdaily.com, every twelve Indian is affected with the diabetes. This work focuses on prediction of diabetes from voice data using Convolution Neural Network (CNN), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) in terms of accuracy (Acc.), precision (Pre.), recall (Re.) and F1-score (F1) values. However, the GRU demonstrated superior performance, achieving classification report Acc. as $\mathbf{9 4. 1 2 \%}$, Pre. as $\mathbf{9 5 \%}$, Re. as $94 \%, \mathrm{~F} 1$ as 94%. This research attempt to find new biomarker for diabetes prediction as voice data. Thereby enhancing early diagnosis and intervention strategies, ultimately contributing to better health outcomes and reducing the socio-economic burden of diabetes in affected communities.