摘要
针对传统的模糊时间序列预测模型中有效论域划分和序列数据模糊化处理存在的不足,提出了一种新的基于累积概率分布的直觉模糊时间序列预测方法.新方法首先通过检验序列数据是否近似服从正态分布,然后利用累积概率分布方法划分论域区间.在此基础上根据直觉模糊时间序列的数据特点,提出一种更具客观性的直觉模糊数隶属度和非隶属度函数的确定方法,同时利用直觉模糊趋势逼近因子代替传统隶属度函数来刻画序列数据对模糊数的隶属情况.然后结合模糊关系矩阵的相关合成运算,合理地选取要考虑的模糊状态,进而对结果进行解模糊化处理.在实验部分,新方法利用典型的阿拉巴马大学的学生招生人数为实验样本数据,将预测结果与传统模型的预测结果进行对比分析,验证新模型的可行性和优越性.
Aiming at overcoming the shortcomings of the traditional fuzzy time series prediction model with effective interval partitioning and sequence data intuitionistic fuzzification pretreatment,a new intuitionistic fuzzy time series prediction method is proposed in this paper.The new method tests whether the sequence data obey the normal distribution,and uses the cumulative probability distribution method to divide the universe of discourse.On this basis,considering the series data characteristics of intuitionistic fuzzy time series,a more objective method to determine the membership and non-membership functions of intuitionistic fuzzy numbers is proposed.At the same time,the intuitionistic fuzzy trend approximation factor is used to describe the membership of sequence data to fuzzy numbers instead of the traditional membership function.Combining with the correlation operation of fuzzy relationship matrix,the fuzzy states to be considered are selected reasonably,and then the results are defuzzy.In the experimental part,the new method uses a typical student enrollment of the University of Alabama as the experimental sample data,and compares the prediction results with those of the traditional model to verify the feasibility and superiority of the new model.
作者
宋敏
柏玉
何文倩
刘士虎
SONG Min;BAI Yu;HE Wen-qian;LIU Shi-hu(School of Mathematics and Computers,Yunnan Minzu University,Kumming 650504,China)
出处
《云南民族大学学报（自然科学版）》
CAS
2022年第6期710-718,共9页
Journal of Yunnan Minzu University:Natural Sciences Edition
基金
国家自然科学基金(61966039)
云南省“兴滇英才支持计划”青年人才专项项目.
关键词
累积概率分布
直觉模糊数
模糊规则
模糊时间序列预测
cumulative probability distribution
intuitionistic fuzzy number
fuzzy rule
fuzzy time series forecasting
