基于区间型数据的金融时间序列预测研究  

Forecasting research of financial time series based on interval data

作　　者：杨威[1,2] 韩艾[2] 汪寿阳[2] Yang Wei Han Ai Wang Shouyang(Institute of Management and Decision, Shanxi University, Taiyuan 030006, China Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100190, China)

机构地区：[1]山西大学管理与决策研究所,山西太原030006 [2]中国科学院数学与系统科学研究院,北京100190

出　　处：《系统工程学报》2016年第6期816-830,共15页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(71501115;71201161);教育部人文社科基金资助项目(14YJC630163)

摘　　要：提出了金融数据预测新方法——区间型时间序列模型,是传统时间序列模型的拓展.在与传统的点值AR模型、VAR模型以及Na?ve模型的比较分析中发现,区间数据模型的预测精度更高,区间高价和区间低价预测误差均较小,而且具有统计显著性.进一步,不同的估计样本量、数据频度以及不同市场特征的区间价格数据对区间模型的稳定性检验再次验证了区间数据模型的可靠性.区间型金融时间序列预测研究不仅为金融问题的定量分析提供了新的视角,也可为政策制定和交易策略实施提供了更丰富的决策参考信息.This paper expands the traditional time series models by proposing a new methodology to forecast the financial data based on the interval time series model. The comparison results of interval prediction accuracy between the interval model with the traditional point-valued AR model, VAR model, and Naive model indicate that the proposed interval forecasting model has a smaller forecasting error than other models in the interval- based low and high price forecasting and that this predictive advantage is statistically significant. In addition, some stability tests based on different estimating samples, data frequency, and index interval price data in different financial markets, prove the reliability of the interval model. This forecasting research of financial interval time series not only provides a new perspective for the quantitative analysis of financial problems, but also provides more decision reference information for making policies and implementing trading strategies.

关 键 词：区间时间序列 区间运算 Dκ-估计 区间预测 

分 类 号：F224[经济管理—国民经济]