An Adaptive Algorithm for Hydrological Time Series Forecasting Based on the Selection of an Analogue-Period

In the paper, an adaptive algorithm for time series forecasting based on the selection of an analogue period is proposed. A distinctive feature of the algorithm is the use of training sample of forecasts for the automatic selection of optimal parameters of its work. The algorithm was employed for prediction of the hydrological time series of inflow to Novosibirsk Reservoir (the Ob River). The efficiency of its use (an increase in the accuracy of forecasts) is demonstrated compared with the basic algorithm.