摘要
GRNN在解决样本量小且噪声较多的问题时,逼近能力、分类能力和学习速度有明显优势.采用主成份分析法的GRNN模型对山东省汽车保有量进行预测,结果表明,该方法具有结构简洁、收敛速度快的优点.与传统方法相比,GRNN模型的预测精度更高,对相关部门的决策具有参考意义.
GRNN has obvious advantages such as better approximation and classification capability,and higher learning speed,in solving the problems with high noise.Case study shows that the principal component analysis of the GRNN model is superior to traditional methods in predicting car ownership in Shandong province.Compared with the traditional methods,GRNN prediction accuracy is higher in decision-making,having the reference value to the relevant departments.
出处
《山东理工大学学报（自然科学版）》
CAS
2011年第4期85-87,共3页
Journal of Shandong University of Technology:Natural Science Edition
基金
山东省经济和信息化委员会技术创新项目(201120201012)
