基于直觉模糊推理的直觉模糊时间序列模型

Intuitionistic fuzzy time series model based on intuitionistic fuzzy reasoning

由于受到模糊集理论的限制,模糊时间序列预测理论在不确定数据集的描述上有失客观,针对这种局限性,提出一种直觉模糊时间序列预测模型。应用模糊聚类算法实现论域的非等分划分;针对直觉模糊时间序列的数据特性,提出一种更具客观性的隶属度和非隶属度函数的确定方法;提出一种基于直觉模糊近似推理的模型预测规则。在Alabama大学入学人数和中国社会消费品零售总额数据集两组数据集上分别与典型方法进行对比实验,结果表明该模型有效提高了预测精度,证明了模型的有效性和优越性。

The objectivity of the fuzzy time series （FTS） forecasting theory in description of uncertain data sets is limited by the fuzzy sets theory. To break this limitation, an intuitionistic FTS （IFTS） forecasting model is built. Firstly, the fuzzy clustering algorithm is used to get unequal domain-dividing intervals. And then a more objective construction method of membership and non-membership functions of intuitionistic fuzzy sets （IFS） is proposed. Thirdly, forecasting rules based on intuitionistic fuzzy approximate reasoning are given. Finally, experiments on datasets of enrollments of the university of Alabama and the total retail sales of social consumer goods of China are carried out which show that the new model improves the prediction accuracy with its validity and superiority.