Abstract
False data injection, sometimes known as FDI, is a common form of assault that is launched against smart grids. The faulty data detection methods that are now in use are unable to detect covert FDI attacks, making it impossible to do so (El Mrabet et al. 67:469–482, 2018). The detection of FDI assaults can be accomplished using a multitude of methods, one of which is machine learning. In this paper, six alternative supervised learning (SVM-FS) hybrid methods, each with its own set of six unique boosted and feature selection (FS) procedures, are analyzed. These strategies are based on SVM-boosting algorithmic frameworks. Using data obtained from the smart grid, an analysis of the adaptability of these approaches is carried out. The accuracy of each detection method in terms of its classification is what is used to evaluate them. When supervised learning and hybrid methods are utilized in a simulation exercise, the performance of the classification techniques that are used to detect FDI attacks is significantly improved.