“COMPANY FORECAST ACCURACY FOR EXPONENTIAL SMOOTHING MODELS OF EARNINGS‐PER‐SHARE DATA FOR FINANCIAL DECISION MAKING”: A COMMENT
指出，在多个预测方法中只保留最准确的一个会丢失有用信息；以两个预测为例，证明即使加入较差的预测方法也能显著提升较优方法的预测效果。
ABSTRACT The practice of abandoning all but the most accurate among a set of alternative forecasting methods is shown to result in the loss of potentially useful information. The particular case of two forecasts is considered in detail. It is demonstrated practically that the inclusion of even a relatively poor forecasting method can enhance a superior one significantly.