东北大学学报:自然科学版 ›› 2019, Vol. 40 ›› Issue (2): 300-304.DOI: 10.12068/j.issn.1005-3026.2019.02.028

• 管理科学 • 上一篇    

关于股票价格的二阶模糊时间序列

刘智1,2, 张铁1, 董莹3, 徐爽爽2   

  1. (1. 东北大学 理学院, 辽宁 沈阳110819; 2. 沈阳工业大学 基础部, 辽宁 辽阳111003; 3. 大连民族大学 理学院, 辽宁 大连116600)
  • 收稿日期:2017-12-06 修回日期:2017-12-06 出版日期:2019-02-15 发布日期:2019-02-12
  • 通讯作者: 刘智
  • 作者简介:刘智(1979-),女,辽宁辽阳人,沈阳工业大学讲师,东北大学博士研究生; 张铁(1956-),男,辽宁沈阳人,东北大学教授,博士生导师.
  • 基金资助:
    国家自然科学基金专项基金天元访问学者项目(11726616); 辽宁省博士科研启动基金资助项目(201501164); 中央高校基本科研业务费专项资金资助项目 (DC201501040).

A Second-Order Fuzzy Time Series Model for Stock Price Analysis

LIU Zhi1,2, ZHANG Tie1, DONG Ying3, XU Shuang-shuang2   

  1. 1. School of Sciences, Northeastern University, Shenyang 110819, China; 2. Department of Basics, Shenyang University of Technology, Liaoyang 111003, China; 3. College of Science, Dalian Nationalities University, Dalian 116600, China.
  • Received:2017-12-06 Revised:2017-12-06 Online:2019-02-15 Published:2019-02-12
  • Contact: LIU Zhi
  • About author:-
  • Supported by:
    -

摘要: 由于股票价格的时间序列具有不确定性,股市的真实模型不容易建立,而模糊时间序列在解决模糊性数据和不确定性数据方面具有较大优势;因此,本文首先将数据进行预处理并改进论域划分的方法,然后利用三角隶属度函数进行数据的模糊化处理,再利用模糊化后的数据建立三层BP神经网络,最后,应用广义的逆模糊数公式将预测模糊集进行逆模糊化,从而得到预测结果.应用本文方法对印度国家银行(SBI)股票价格和Alabama 大学的入学人数进行预测,预测结果精度较高.

关键词: 二阶模糊时间序列, BP神经网络, 逆模糊数, 模糊时间序列, 股票价格

Abstract: It is difficult to model stock market because of its uncertainty, while fuzzy time series has its advantages in dealing with fuzzy and uncertainty data. Accordingly, the data was first preprocessed and a new way to divide the universe of discourse was given, after which the data was fuzzified using the triangular membership function, and a three-layer BP neural network was then established according to the fuzzified data. Finally, the generalized inverse fuzzy number formula was used to defuzzify the fuzzy relation, with the prediction results obtained. The method was used for predicting the stock price of State Bank of India(SBI)and the enrollment of the University of Alabama, and the results showed that the prediction accuracy is higher than that of the related previous methods.

Key words: second-order fuzzy time series, BP neural network, inverse fuzzy number, fuzzy time series, stock price

中图分类号: