Abstract
The objective of this study is to apply machine learning algorithms to automatically detect accidents in heating networks. The study uses the Orange data mining environment, which allows for a clear and intuitive implementation of the data analysis stages. The data set includes parameters characterizing the state of the system, such as temperature, pressure, coolant flow rate, and time. Random forests, logistic regression, and k-nearest neighbors (k-NN) methods for anomaly detection were used to classify accidents. The models were trained and tested on a demo dataset. The results showed that the proposed methods provide high classification accuracy, and the error matrices and ROC curves confirm the effectiveness of the models. The results demonstrate the potential of machine learning to improve the reliability of heat supply systems. The practical significance of this study lies in the possibility of integrating such systems into the existing monitoring infrastructure, which will allow for quick detection of faults, accidents, and reduction of maintenance costs.