摘要
针对最小二乘支持向量机(LS_ SVM)不需要指定逼近精度ε的特点,比较了LS_ SVM与SVM两种方法利用生产数据为青霉素发酵过程建立的数学模型,改进型GA分别为LS_ SVM和SVM选择参数值.实验证明:LS_ SVM建立的模型具有较高的拟合精度和泛化能力.如果ε过大时,SVM建立的模型的拟合精度和泛化能力不高;当ε过小时,模型的拟合精度和泛化能力较高,但耗时多.因此,LS_SVM更适合为发酵过程建模.
The SVM needs to use approximation accuracy ε,however the LS_ SVM doesn＇t need ε.According to this characteristics,the paper studied the fitting and generalization capabilities of models that LS_ SVM and SVM established for the penicillin fermentation process respectively.An improved GA selected the parameter values for LS_ SVM and SVM respectively.The experiment shows that the model based on LS-SVM possesses the strong capabilities of fitting and generalization.If ε is too large,the capabilities of fitting and generalization of model based on SVM are not high;if ε is too small,the capabilities of fitting and generalization are relatively high,but the modeling process demands long time.Therfore,the LS_ SVM is more suitable for modeling in fermentation processes.
出处
《北京工业大学学报》
EI
CAS
CSCD
北大核心
2010年第1期7-12,共6页
Journal of Beijing University of Technology
基金
国家自然科学基金资助项目(60704036)
北京工业大学青年基金资助项目(97002011200701)
北京工业大学博士科研启动基金资助项目(52002011200707)
