JUDGMENTAL BOOTSTRAPPING: INFERRING EXPERTS' RULES FOR FORECASTING

摘要: Judgmental bootstrapping is a type of expert system. It translates an expert’s rules into quantitative model by regressing the forecasts against information that he used. Bootstrapping models apply consistently, and many studies have shown decisions predictions from are similar to those experts. Three showed improved quality production in companies. To date, research on forecasting with judgmental has been restricted primarily cross-sectional data, not time-series data. Studies psychology, education, personnel, marketing, finance were more accurate than made experts using unaided judgment. They for eight eleven comparisons, less one, there two ties. The gains accuracy generally substantial. can be useful when historical data variable forecast lacking or poor quality; otherwise, econometric should most appropriate complex situations, where judgments unreliable, experts’ some validity. When needed, cost-effective. If differ greatly expertise, draw upon best aids learning; it help identify biases way make predictions, reveal how predictions. Finally, offers possibility conducting “experiments” causal variables varied over time. Thus, serve as supplement models.