首页 | 本学科首页   官方微博 | 高级检索  
     检索      

优化的混合核函数相关向量机在发酵过程建模中的应用
引用本文:余荣荣,周家旺.优化的混合核函数相关向量机在发酵过程建模中的应用[J].云南民族大学学报(自然科学版),2014,23(5):379-382.
作者姓名:余荣荣  周家旺
作者单位:1. 安徽师范大学数学计算机科学学院,安徽芜湖,241000
2. 武汉理工大学机电工程学院,湖北武汉,430070
基金项目:国家自然科学基金天元青年基金
摘    要:针对相关向量机中核函数选择问题进行研究,提出一种局部性高斯核函数和全局性sigmoid核函数结合的相关向量机建模方法.采用差分进化算法对混合核函数构成的相关向量机的参数进行优化,提高了模型的预测精度.将参数优化后的混合核函数相关向量机用于青霉素发酵过程的软测量建模,得到比单一核函数所建模型更优的预测精度.

关 键 词:差分进化算法  混合核函数  相关向量机  建模

Application of the relevance vector machine with the optimized combined kernel function to the fermentation process modeling
YU Rong-rong,ZHOU Jia-wang.Application of the relevance vector machine with the optimized combined kernel function to the fermentation process modeling[J].Journal of Yunnan Nationalities University:Natural Sciences Edition,2014,23(5):379-382.
Authors:YU Rong-rong  ZHOU Jia-wang
Institution:YU Rong-rong;ZHOU Jia-wang;School of Mathematics and Computer Science,Anhui Normal University;School of Mechanical and Electronic Engineering,Wuhan University of Technology;
Abstract:Concentrating on the selection of the kernel function of relevance vector machine,this research has combined the locality- sensitive Gaussian kernel with the global sigmoid kernel,which has been used in the relevance vector machine modeling. It uses the differential evolution algorithm to optimize the parameters which are formed in the relevance vector machine with the combined kernel function. The optimized RVM with combined kernel function has been used in penicillin fermentation process modeling,and a much more accurate prediction than the single kernel function has been obtained.
Keywords:differential evolution algorithm  combined kernel function  relevance vector machine  model
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号