Remaining Useful Life Improvement for Electrical Machines Using Ensemble Learning Technique

The Remaining useful prediction is an indictor to know the electrical machine health in advance. From the machine learning point of view, the problem of remaining useful life prediction is a multi-class regression. To clarify the prediction, a model using machine learning will predict the health of electrical machine in terms of remaining life based on historical observations of the machinery under scan. In the dataset related to machine health, XG Boost is chosen because of the advantage of handling large dataset effectively so that it provides better performance. XG Boost is a machine learning method that is widely used for it to overcome problem related to model performance in terms of a remaining useful life prediction of the machine. This paper analyzed the performance of the XG Boost model in handling the model accuracy for prediction in machine health. The simulations show that the XG Boost model performs better in terms of model accuracy as comparable to different models of the supervised learning class.