Today, the digitalization of the fuel and energy sector all over the world has led to the active and almost universal digitization and platform service implementation, and in most developed countries it is even included in priority national development programs. Such an active transformation of the industry revealed new problems, among which one of the main ones were the problems of the continuous growth of data volumes and the need for new approaches to their processing and analysis. The authors of this paper have quite a lot of experience in the development and implementation of decision support systems based on machine learning algorithmsin various tasks of the power industry, and in the presented paper they tried to aggregate all their practical experience to analyze the main errors and the consequences of their influence on the results of the operation of such systems inthe power industry. The paper also describes examples of the interpretation of the results, both in terms of data processing, and more importantly, in terms of their interpretation for the power industry.