Forecasting the Probability of a Power Outage Using Binary Classification

This paper considers the use of machine learning methods to forecast the probability of a power outage at a substation. The data on power outages for 2 years were analyzed. They contained information about the substation where the emergency power outage occurred and information about the weather. A solution is proposed for forecasting the probability of a power outage caused by weather factors. The features for compiling training and test samples are determined, an analysis is carried out to select the appropriate algorithm for training the model. A model was trained for further use in forecasting the probability of outages.