Qualitative Forecast Evaluation: A Comparison of Two Performance Measures
比较了两种定性预测准确度度量方法，发现Henriksson-Merton的概率度量比Naik-Leuthold度量能提供更多、更准确的模型定性预测能力信息，对农业产销者和金融决策者评估预测有用。
Abstract Agricultural producers and marketers often have access to several sets of forecasts on the same series. Decision makers must be able to evaluate the relative accuracy of the forecasts. The ability to evaluate whether a series will move up or down is important for series related to futures, options, other asset prices, and situations where processes can be modified according to changing economic conditions. A measure of qualitative forecast accuracy from Naik and Leuthold is compared to a probability‐based measure developed by Henriksson and Merton. The Henriksson‐Merton measure is shown to provide additional, and more accurate, information concerning a model's qualitative forecasting ability.