This paper evaluates the quality of forecasting regional inflation rates using machine learning methods in the case of the Siberia Macroregion and Siberian regions. At the first stage of our study, we forecast regional inflation rates for various periods using several machine learning and benchmarking methods. At the second stage, we combine the forecasts with machine learning methods and weight them based on the quality metrics obtained. In the final part of this paper, we compare the obtained quality metrics with our benchmarks and confirm the stability of the results achieved using the Diebold - Mariano test. Based on the results of our study, we conclude that the quality of forecasting inflation rates in the Siberia Macroregion and Siberian regions using machine learning methods is comparable with traditional econometric methods. At the same time, it is necessary to assess the quality of forecasting through machine learning methods for each region in advance to determine whether it makes sense to use them over traditional econometric methods. For most Siberian regions and the Siberia Macroregion as a whole, machine learning methods work better than benchmarks for periods longer than a year, in contrast to forecasts for 1-3 quarters ahead. Forecasting with combining machine learning models is in most cases preferable to any one.