本文已被：浏览 1320次 下载 2969次
Received:May 25, 2014 Revised:July 30, 2014
Received:May 25, 2014 Revised:July 30, 2014
中文摘要: 在时序数据的建模与分析过程中, 趋势项往往是选用单一形式的函数来进行拟合的, 这对于长时序数据而言, 其拟合精度难以提高. 为此, 设计实现了一种分段拟合算法, 该算法通过计算时序数据的各阶趋势导数, 有效地把被拟合的时序数据划分成多个不同模式的子序列, 在此基础上, 选用合适的拟合函数对各个子序列的趋势项进行拟合, 并基于最小二乘法求出相应的拟合参数向量. 实验验证了算法的有效性和先进性.
Abstract:Trend components often choose single form functions for fitting in the process of timing data modeling and analysis. Therefore, the fitting accuracy is difficult to improve for long timing data. As regard to this important problem, a piecewise fitting algorithm is implemented. Through the calculation of each order derivative of the timing data, the algorithm effectively divides the fitted timing data into a number of different patterns of subsequence. On this basis, corresponding fitting functions are chosen to fit each subsequence trend components respectively. Test results show that the new algorithm is feasible and effective.
文章编号： 中图分类号： 文献标志码：
基金项目:佛山职业技术学院校级重点科研课题(2011KY006)
| Author Name | Affiliation |
| HUANG Xiong-Bo | Department of Electronic and Information Engineering, Foshan Professional Technical College, Foshan 528137, China |
| Author Name | Affiliation |
| HUANG Xiong-Bo | Department of Electronic and Information Engineering, Foshan Professional Technical College, Foshan 528137, China |
引用文本：
黄雄波.时序数据趋势项的分段拟合.计算机系统应用,2015,24(2):174-179
HUANG Xiong-Bo.Piecewise Fitting of Trend Component from Timing Data.COMPUTER SYSTEMS APPLICATIONS,2015,24(2):174-179
黄雄波.时序数据趋势项的分段拟合.计算机系统应用,2015,24(2):174-179
HUANG Xiong-Bo.Piecewise Fitting of Trend Component from Timing Data.COMPUTER SYSTEMS APPLICATIONS,2015,24(2):174-179
