Author Materials and methods Dataset(s) used Evaluative measures Findings Kaur et al. [22 ] CI-PDF, a cloud IoT-based diabetes prediction platform, was introduced PIDD Accuracy, sensitivity, and specificity Achieved 94.5% of prediction accuracy by combining DT and NN approaches Priyadarshini et al. [15 ] Introduced DeepFog, a deep neural architecture based on fog computing for forecasting stress, diabetes, and hypertension attacks Stress, T2D, hypertension datasets Accuracy, precision, recall, and Achieved a superior and competitive method in comparison to others Fernández-Caramés and Fraga-Lamas [21 ] Introduced an IoT CGM-based system to remotely monitor patients by accessing blood sugar samples obtained and to notify them if a problematic scenario is discovered Not mentioned Prediction accuracy This study claimed to be able to give a population with transparent and trustworthy blood sugar data quickly, easily, and affordably Barik et al. [23 ] Introduced the FogLearn framework for PIDD Accuracy As a consequence, fog computing has great potential for extensive data analysis in medicine and geography Fernández-Caramés et al. [6 ] Designed and built a solution that allows continuous commercial CGMs to monitor patients remotely and warn them about their problems Not mentioned Prediction accuracy The authors claimed to have developed a better technique for diagnosing patients’ illnesses remotely in real time Gia et al. [24 ] A fog-based system for remote health monitoring and fall detection was introduced Not mentioned Energy efficiency The suggested procedure is precise, and the worn sensor node saves energy Devarajan et al. [25 ] To manage blood glucose levels, propose an energy-efficient fog-assisted healthcare system PAMAP2 and PIDD Energy efficiency, accuracy, computational complexity, and latency As a consequence, the fog over cloud computing has improved bandwidth efficiency, latency, and classification accuracy Abdel-Basset et al. [26 ] To diagnose and monitor type 2 diabetes patients, a new framework based on computer-assisted diagnostics and the IoTs was proposed Personally collected data Accuracy and execution time The suggested algorithms were shown to be valid and resilient Haq et al. [27 ] Developed a filter method based on the ID3-DT model Clinical data Accuracy and computation time Found that the decision tree algorithm based on selected features improves the classifier performance Kumari et al. [28 ] Proposed an ensemble voting classifier that employs the ensemble of three ML algorithms, viz., RF, LR, and NB PIDD Accuracy, precision, recall, and Achieved comparatively enhanced results on binary classifications Geetha and Prasad [29 ] Proposed T2DDP is a hybrid model that uses supervised classification algorithms like NB and ensemble algorithms like bagging with RF and AdaBoost for DT to help physicians properly treat diabetic patients PIDD Accuracy, precision, recall, and It was discovered that the predicted outcome would be sent to the patient’s mobile phone at an early stage, allowing them to make quick judgments concerning the health risk Shynu et al. [30 ] Introduced efficient blockchain-based safe healthcare services for illness prediction in fog computing PIDD and CHDD Accuracy, purity, and NMI In comparison to existing approaches, the suggested work efficiently clusters and predicts illness Singh et al. [31 ] eDiaPredict, an ensemble-based system that uses XGBoost, RF, SVM, NN, and DT to forecast diabetes status in patients, was introduced PIDD Accuracy, GI sensitivity, AUC, specificity, precision, AUCH, MER, and MWC The proposed model can provide patients with a practical and precise prediction of diabetes based on glucose concentrations Rajput et al. [32 ] Proposed a reference model for aiding rural people in India who are suffering from diabetes PIDD Accuracy, MCR, recall, precision, prevalence, and The authors stated that they were able to increase patient communication and engagement
