Analyzing the application of artificial neural networks to assess the technical condition of 6-10 KV power transformers

This paper compares the effectiveness of the k-nearest neighbours Regressor machine learning model and the k-nearest neighbours Classifier model in the task of predicting the state of transformer equipment using data that has been captured by a transformer monitoring system. Subject of research: power transformers. Purpose of research: identifying the best model in the task of classifying the state of transformer equipment. Object of research: transformer equipment diagnostic system. Methods of research: analysis, computer modelling. Main results of research: data on the effectiveness of models with different parameters were obtained, but all regression models showed a worse result than the classifier model.