多核小波支持向量机在Carrousel氧化沟系统的应用  

Application of Wavelet Multi Kernel Learning on Carrousel Oxidation Ditch System

作　　者：何渊淘[1] 刘超慧[1] 

机构地区：[1]郑州航空工业管理学院计算机科学与应用系,郑州450000

出　　处：《计算机系统应用》2013年第10期203-205,197,共4页Computer Systems & Applications

基　　金：国家自然科学基金(41171341)

摘　　要：Carrousel氧化沟广泛应用于城市污水处理,但污水处理的效果受到水质和环境因素影响很大,难以建立精确的预测模型.现有的机器学习方法普遍预测效果较差,为了准确预报污水处理的效果,本文采用多核小波支持向量机进行建模,实验表明该方法提高了预报的精确度,适合用于氧化沟系统的实时在线预测.Carrousel oxidation ditch has been widely used in sewage disposal system, however the result of sewage disposal was affected by water quality and environmental factors, so it is difficult to build a precise prediction model. The existing method of machine learning algorithm usually gets a poor result in prediction. In order to precisely predict the result, this essay uses the multi-kernel wavelet support vector machine when build a model. The outcome of this experiment demonstrates that the new method improve the degree of definition in forecasting, and it is suitable for actually online prediction.

关 键 词：多核支持向量机 SILP算法 SimpleMKL算法 小波核函数 氧化沟系统 

分 类 号：X703.3[环境科学与工程—环境工程] TP18[自动化与计算机技术—控制理论与控制工程]