基于改进人工鱼群算法的模糊时间序列模型

Fuzzy Time Series Model Based on Modified Artificial Fish School Algorithm

近年来,经过人们对模糊时间序列模型大量的研究后,发现模糊区间的划分以及模糊关系的构建是影响模糊时间序列模型预测精度最关键的两个因素。文章针对目前模糊时间序列模型在论域划分中存在的问题,提出将人工鱼群算法（AFSA）用于模糊时间序列模型的论域划分中;并利用最小均方误差（MSE）确定最优的预测结果;最后为了显示所提出方法的有效性,将该方法用于Alabama大学注册人数的预测,结果显示该方法的预测精度更高。

In recent years, as the fuzzy time series model has been studied for a long time, it inds that the division of fuzzy interval and the structure of fuzzy relation are two most critical factors affecting the prediction accuracy of the fuzzy time series model. To solve the domain partition in fuzzy time series model,this paper proposes artificial fish school algorithm（AFSA） to tackle the domain partition problem in fuzzy time series model; and utilizes minimum mean square error（MSE） to determine the optimal forecast results.Finally, in order to show the effectiveness of the proposed method, the paper applies this method to forecast the enrollment of Alabama University. The results show a higher prediction accuracy of the method.