Hybrid Prediction Method Using Experts and Models

The article considers a hybrid forecasting model, which first considers "necessary" and "sufficient" forecasting models in the decision-making process, then selects the best pairs or triples of models, and finally attract experts. It discusses how to work with the hybrid model when to involve experts and according to what criteria. Algorithms for choosing a new forecasting model, as well as algorithms for the need to change an expert, have been developed. The choice of experts is based on the predictions they make, not just their competence. The result of the forecast is the best pair of intersections of expert and model forecasts. A specific example illustrating the presented algorithms is considered.