GM(1,1)模型的性质及改进  

Properties and improvement of GM (1,1) models

作　　者：潘澔 高尚[2] PAN Hao;GAO Shang(Suzhou Institute of Construction&Communications,Suzhou 215104,Jiangsu,China;School of Computer Science and Engineering,Jiangsu University of Science and Technology,Zhenjiang 212003,Jiangsu,China)

机构地区：[1]苏州建设交通高等职业技术学校,江苏苏州215104 [2]江苏科技大学计算机科学与工程学院,江苏镇江212003

出　　处：《山东大学学报（理学版）》2021年第11期38-42,60,共6页Journal of Shandong University(Natural Science)

摘　　要：在对GM(1,1)模型进行分析的基础上,经过理论推导,得出了初始数对预测没有影响的结论,对GM(1,1)模型进行改进,给出了GM(1,1)模型Ⅰ。当向原始序列添加相同的数字时,预测值将更改,由此提出了GM(1,1)模型Ⅱ,利用粒子群算法,得到最佳的增加量。仿真结果表明,GM(1,1)模型Ⅰ和模型Ⅱ具有较高的精度。Based on theoretical analysis of GM(1,1) model, the conclusion, which the initiative number has no effect on the prediction, is got. GM(1,1) model is improved and GM(1,1) model I is given. When add an identical number to the original series, the forecast values will change. GM(1, 1) model Ⅱ is given, and using particle swarm algorithm, the best increase is got. Simulation results show that the improved GM(1, 1) model Ⅰ and GM(1, 1) model Ⅱ have higher accuracy.

关 键 词：GM(1 1)模型 简化 精度 粒子群优化算法 

分 类 号：N941.5[自然科学总论—系统科学]