参数自适应的长期IFTS预测算法

Method of long-term IFTS forecasting based on parameter adaptation

针对模糊时间序列预测理论多局限于短期时间范围预测以及对不确定数据集模糊变化趋势描述和论域区间划分研究不足的问题,构建了参数自适应的长期直觉模糊时间序列预测模型。新模型通过引入滑动窗口机制和参数自适应的直觉模糊C均值聚类算法优化论域区间划分,利用矢量预测技术解决时间序列长期范围预测误差积累的问题,有效地提高了复杂环境下时间序列长期趋势预测的精度,扩展了直觉模糊时间序列预测理论的应用范围。最后,通过典型实例验证了该方法的有效性和优越性。

To overcome the problem that the fuzzy time series forecasting theory is mostly limited to short- term forecasting and the insufficiency of the description for the fuzzy trend of uncertain data sets and the partition intervals, the long-term intuitionistic fuzzy time series forecasting model based on parameter adaptation approach is advanced. The new model optimizes the domain dividing intervat with the sliding window scheme and intuitionistie fuzzy C-means clustering algorithm based on parameter adaptation. Meanwhile, the long term time series forecasting issue is solved efficiently by introducing the vector forecasting technique, thus impoving the prediction accuracy of long-term time series in the complex environment. And the intuitionistic fuzzy time series forecasting theory application is greatly extended. Finally, the classical instances prove the validity and superb ority of the proposed method.