论文部分内容阅读
源于与决策分析的相关性,预测组合已经逐渐形成了一个重要的研究领域。为此,本文引进EWMA技术对预测组合权重更新的过程进行控制,从而提出一种能够应用于实际且简单有效的EWMA赋权方法。这种赋权方法能够确定预测组合权重应该何时更新,而不是机械地更新预测组合权重。本文额外针对各种赋权方法在旅游预测组合模型中的预测性能(全面预测性能和总均方根误差)和预测效率(权重更新频率)进行了经验评估。结果显示:EWMA赋权方法的预测性能优于传统的赋权方法,并与CUSUM赋权方法相似,同时该赋权方法获得了最小的权重更新频率。综合考虑预测性能和预测效率,EWMA赋权方法相比于其他赋权方法在旅游实际应用过程中更具优势。
Stems from the relevance of decision analysis, forecasting portfolio has gradually formed an important area of research. Therefore, this paper introduces the EWMA technology to control the process of forecasting portfolio weight update, so as to propose a practical and simple and effective EWMA weighting method. This method of weighting can determine when the forecasting portfolio weights should be updated, rather than mechanically updating the forecasting portfolio weights. This paper also empirically evaluates the predictive performance (overall prediction performance and total root-mean-square error) and the prediction efficiency (weight update frequency) of various weighting methods in the travel forecasting portfolio model. The results show that the EWMA weighting method is better than the traditional weighting method in predicting performance and is similar to the CUSUM weighting method. At the same time, the weighting method obtains the least weight updating frequency. Considering the predictive performance and predictive efficiency, EWMA-based weighting method has more advantages than other weighting methods in the practical application of tourism.