摘要
指出了用叶绿素a的浓度估计海洋初级生产力的重要作用;分析了目前采用的浓度反演方法的不足;尝试将基于统计学习理论的最小二乘支持向量机用于浓度反演,SeaBAM的数据实验结果表明该方法可以获得更高的反演精度;可以有效避免过学习的情况出现;不像神经网络那样需要确定网络结构。
That chlorophyll-a is an important factor when estimating ocean＇s primitive productivity. The shortcomings of current methods used to retrieve chlorophyll-a concentration were analyzed. For the first time, least squares support vector machine based on statistic learning theory was used to do this work Experiments based on this method. SeaBAM data shows that using SVM method can receive the lowest error, avoid over fitting, and not need to design net-structure.
出处
《计算机应用》
CSCD
北大核心
2005年第10期2398-2400,2409,共4页
journal of Computer Applications
基金
教育部博士点基金资助项目(20040613013)
关键词
叶绿素A
浓度反演
统计学习理论
机器学习
支持向量机
chlorophyll a
concentration retrieve
statistic learning theory
machine learning
Support Vector Machine（SVM）
