Power systems should be capable of delivering uninterrupted, reliable power to their customers economically. The increase in the frequency of weather-related events like cyclones, and earthquakes is creating power disruptions as well as huge economical losses. Accurate prediction of power outages due to these high-impact low-frequency (HILF) events becomes imperative for reducing its impact on the grid. The accurate prediction of outages will help the utilities to reduce the impact of these events on the grid by taking appropriate measures. In this paper, the prediction of the outage of power in the transmission line is implemented with the help of machine learning algorithms. There will be two states for the status of a transmission line in response to the extreme event. The two states are damaged or survived. The survived line will continue the power flow, whereas, the damaged lines will result in a power outage. Orissa a state in India is often affected by cyclones. It has a 485 km long coastline. Cyclone Phailin and Fani hit the coastal areas in 2013 and 2020 respectively, resulting in damage to power system infrastructures. The state of transmission towers/lines has been predicted using two machine learning algorithms logistic regression, Incremental Principal Component Analysis (IPCA), and Support Vector Machine (SVM). The performance of both the proposed model algorithms has been calculated and accuracy was also found to be good. The performance of the IPCA algorithm is validated using k-fold cross-validation. The decision boundary that divides the above-mentioned classes is plotted using all three algorithms.