摘要：Organic farmers, wholesalers, and retailers need reliable price forecasts to improve their decision- making practices. This paper presents a methodology and protocol to select the best-performing method from several time and frequency domain candidates. Weekly farmgate prices for organic fresh produce are used. Forecasting methods are evaluated on the basis of an aggregate accuracy measure and several out-of-sample predictive ability tests. Combining forecasts to improve on individual forecasts is investigated.
