相关向量机方法预测药膜的释药规律  

Relevance Vector Machine Applied to the Prediction of Release Behavior of Membranes

作　　者：詹环[1] 权维燕[1] 李程鹏[1] 李思东[1] 杨磊[1] 蔡鹰[1] 邱佩东[1] 

机构地区：[1]广东海洋大学理学院,广东湛江524088

出　　处：《广东化工》2008年第6期93-95,117,共4页Guangdong Chemical Industry

基　　金：广东省科技计划重大项目(2003A2030504)

摘　　要：鉴于药膜释药过程的高度非线性,采用相关向量机方法为药膜的释药过程建模,得到药膜累积释放率与时间的定量关系。通过优化建模参数,所建相关向量机模型具有较高的拟合能力,且预测误差小,稳健性好。研究结果表明,相关向量机建立的模型可对释放进行有效预测。Considering the difficulties of the problem of severe non-linearity in the release process of the medicinal membranes, Relevance Vector Machine （RVM） was applied to set up the model of the release process of the medicinal membranes, and the quantitative relationship between accumulated release rate of the membranes and time was found. By optimizing the model parameters, the obtained RVM model had high fitting abilities, less prediction errors and less standard deviation of prediction errors. The results indicated that the model set up by RVM could predict the release successfully.

关 键 词：药膜 释放 相关向量机 

分 类 号：O211.67[理学—概率论与数理统计] TP338[理学—数学]