摘要
时间序列预测法在各种基于时态数据库的计算中有着广泛的应用前景。文中介绍了时间序列预测法中的单指数平滑、双指数平滑和三指数平滑三种指数平滑预测方法,不同的预测方法适合于对不同时间特性的数据、平稳性数据、趋势性数据或季节波动性数据进行预测,使用相应的预测方法均达到很好的平滑效果。同时还介绍了如何应用IGS算法对指数平滑的参数进行优化,从而得到更好的平滑效果和预测结果,使之在社会实际当中发挥更好的作用。
The time series prediction methods have extensive application future in many kinds of computations based on the tense databases.In this paper, the single exponent smoothing, the double exponent smoothing, and triple exponent smoothing are introduced. Different prediction methods are adapted to data with different special properties. And how to use the algorithm IGS to optimize the parameters of the exponent smoothing, through which better smoothing results and prediction results are obtained, is described.And it can bring into play better effect in social practice.
出处
《微机发展》
2005年第3期1-3,41,共4页
Microcomputer Development
基金
上海科委发展基金资助项目(025115036)
关键词
单指数平滑
双指数平滑
三指数平滑
IGS算法
single exponent smoothing
double exponent smoothing
triple exponent smoothing
IGS algorithm
