A Comparative Analysis of Machine Learning Algorithms for Predictive Maintenance in Electrical Systems

This research explores machine learning calculations for prescient support in electrical frameworks, centering on Support Vector Machines (SVM), Random Forests, Recurrent Neural Networks (RNN), and K-Nearest Neighbors (KNN). Through a comprehensive comparative examination, the study assesses these calculations on a different dataset, considering both classification and relapse errands. Results demonstrate that Random Forest beats other calculations, accomplishing an amazing classification precision of 92%, exactness of 93%, review of 91%, and F1-score of 92%. In relapse assignments, Random Forest shows the most reduced cruel outright mistake (MAE) of 8.5 and cruel squared mistake (MSE) of 120.3, displaying its vigor in anticipating remaining valuable life. SVM illustrates an adjusted execution with a precision of 85% and competitive accuracy and review. RNN, specialized for successive information, accomplishes a precision of 88% and competes closely with SVM. KNN, whereas instinctive, appears to lower execution with a precision of 78%. These discoveries give profitable insights for professionals, directing the choice of machine-learning algorithms custom-made to specific prescient upkeep scenarios in electrical frameworks.