Enhancing Cybersecurity for Diabetes Patient Data in Cloud-Based IoT Using Advanced Machine Learning Algorithms

With an end goal to upgrade the network protection of diabetes patient data in a Cloud-Based Web of Things (IoT) system, this review utilizes progressed computer-based intelligence calculations. Four unique models - Backing Vector Machines (SVM), Random Forests, Neural Networks, and K-nearest Neighbors (KNN) - were deliberately assessed for their adequacy in shielding touchy medical services information. The dataset, comprising different profiles of diabetes patients, went through thorough preprocessing and tests were led with an emphasis on measurements like exactness, accuracy, review, f1-score, and the region under the recipient working trademark (ROC) bend. The discoveries uncovered that Brain Organizations arose as the top entertainer, reinforced by an amazing exactness pace of 94%, accuracy pace of 92%, review pace of 94%, f1-score of 93% and an extraordinary AUC-ROC worth of 97%. Irregular Woodlands showed vigor with an exactness pace of more than 93%, accuracy pace of more than 93%, a review rate generally around 92% almost indistinguishable from F-measure 98% as well as an exceptional AUC-ROC esteem of around 95%. Support Vector Machines showed high precision (92%) and exactness (94%) yet achieved computational power. K-Nearest Neighbors showed effortlessness with an exactness of 88%, accuracy of 89%, review of 85%, F1-score of 87%, and an AUC-ROC worth of 91%. The comprehensive understanding of each algorithm's advantages and disadvantages provided by these findings enables the strategic selection of an algorithmic strategy tailored to the particular requirements of securing diabetes patient data in a Cloud-Based IoT environment.