灰色GM(1,1)档案模型参数估计优化算法  

Optimization Algorithm for Parameter Estimation of Grey GM(1,1)Archives Model

作　　者：郑蕾 ZHENG Lei(Department of Science and Technology,Qingdao University,Qingdao 266071,China)

机构地区：[1]青岛大学科技处,山东青岛266071

出　　处：《青岛大学学报（工程技术版）》2024年第3期55-57,共3页Journal of Qingdao University(Engineering & Technology Edition)

基　　金：青岛市社科规划资助项目(QDSKL2201067)。

摘　　要：针对最小二乘法在处理总体特征不明显的数据时效果欠佳的问题,提出了一种适用于“贫信息,小样本”的预测方法,采用灰色GM(1,1)模型对数据序列进行归纳、总结和提炼,构建相应的预测模型。新的参数估算方法将初始误差的绝对值和最小化,将参数估算转化为目标规划问题,利用目标规划法,在不依赖于数据整体特性的前提下,有效降低异常值,简化计算过程,提高预测精度。Aiming at the problem that the least square method is not effective in dealing with the data with unobvious overall characteristics,a prediction method suitable for“poor information,small sample”is proposed.The grey GM(1,1)model is used to summarize and refine the data series,and the corresponding prediction model is constructed.The new parameter estimation method minimizes the absolute sum of the initial error and turns the parameter estimation into a goal programming problem.By using the goal programming method,the outliers are effectively reduced,the calculation process is simplified and the prediction accuracy is improved without relying on the overall characteristics of the data.

关 键 词：灰色GM(1 1)模型 预测模型 参数估计方法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]