The modern power system requires real-time monitoring and fast control to be protected from faults on transmission lines. The detection and classification of faulty conditions in power systems is a task of crucial importance for reliable operation. The traditional fault diagnosis methods rely on the manual feature extraction of engineers with prior knowledge that has been proposed by several researchers for fault detection and classification. It is highly necessary to identify faults in any analog circuit to ensure the circuit's reliability. Early diagnosis of faults in a circuit can help to maintain the system significantly by avoiding potentially harmful damage from the fault. Automatically and accurately identifying the incipient micro-fault in the power system, especially for fault orientations and severity degree, is still a significant challenge in the field of intelligent fault diagnosis. Intelligent fault diagnosis methods based on machine learning become a research hotspot in the fault diagnosis field. In this paper, various machine learning algorithms are discussed. © 2020 Elsevier B.V., All rights reserved.