摘要
组合预测作为一种改善单项模型预测性能的有效策略已得到广泛应用,但如何科学构建组合预测模型仍值得深入研究。文章以提高组合模型预测精度为出发点,首先明确组合预测模型的普遍性优势,然后围绕单项模型筛选策略和组合权重确定方法展开文献述评,明晰组合预测模型构建方法,最后通过对比国内外相关研究,提出该领域未来继续研究的问题与方向。
Combination forecast has been widely used as an effective strategy to improve the prediction performance of individual model, but how to construct an integrated forecast model still needs further study. This paper firstly clarifies the universality advantage of combined forecasting model, and then performs the literature reviews around the single model screening strategy and combined weight determination method, clarifying the method of combination prediction model construction. Finally, the paper presents the future research items and directions in this field by comparing the domestic and foreign related researches.
作者
凌立文
张大斌
Ling Liwen;Zhang Dabin(College of Mathematics and Informatics,South China Agricuhural University,Guangzhou 510642,China;Faculty of Megadata and Computing,Guangdong Baiyun University,Guangzhou 510450,China)
出处
《统计与决策》
CSSCI
北大核心
2019年第1期18-23,共6页
Statistics & Decision
基金
教育部人文社会科学研究规划项目(16YJA630073)
广东省哲学社会科学规划项目(GD15CGL16)
广东省自然科学基金资助项目(2016A030313402)
关键词
组合预测
模型筛选
组合权重
combination forecast
model screening
combination weights