COMPARING FORECAST ACCURACY FOR EXPONENTIAL SMOOTHING MODELS OF EARNINGS‐PER‐SHARE DATA FOR FINANCIAL DECISION MAKING
通过实验比较了多种外推预测模型在预测每股收益（EPS）上的表现，发现Holt-Winter模型在20年随机样本中预测准确且成本效益高，对投资组合分析和财务管理有用。
ABSTRACT This paper relates recent research in predicting accounting earnings per share (EPS) to an experiment comparing the performance of extrapolative forecasting models. The paper points out the usefulness of the results to decision‐making processes such as those used in portfolio analysis or financial management. The statistical results of the experiment point to the usefulness of the Holt‐Winter (HW) model in predicting EPS for a random sample of firms over a 20‐year horizon. For short‐term forecasting, the HW model provides relatively accurate forecasts in comparison to other methods used. HW is likely to be a costeffective alternative to more time‐consuming and expensive techniques.