基于模糊最近邻聚类学习算法的海水藻类繁殖状态预测  

State prediction of algae reproduction in seawater based on learning algorithm of fuzzy nearest neighbor clustering

作　　者：张颖[1] 李鹏[2] 邬益川[2] 

机构地区：[1]上海海事大学信息工程学院,上海200135 [2]国家海洋局东海标准计量中心,上海201306

出　　处：《东南大学学报（自然科学版）》2011年第B09期32-35,共4页Journal of Southeast University：Natural Science Edition

基　　金：上海海事大学科研基金资助项目(20110036);上海市教育委员会重点学科建设资助项目(J50602)

摘　　要：为了预报海洋赤潮等灾害的发生,需要监测海水中藻类的繁殖状况.通过建立针对海水中叶绿素a浓度状态的预测模型间接预测海水中藻类的繁殖生长状况.运用基于数据的模糊最近邻聚类学习算法对采样数据进行聚类处理,基于最优模糊逻辑系统建立了针对海水叶绿素a浓度状态的预测模型.根据采样数据的特点和降低模型阶次的实际需求,设计了改进的模糊最近邻聚类学习算法,并应用于叶绿素a浓度状态的预测.实验结果表明,该方法可有效提取样本数据的特征,降低数据维数,提高算法学习的收敛速度.改进的模糊最近邻聚类学习算法降低了预测模型构造的复杂度,使模型具有较好的泛化性能,能够有效预测海水中藻类的繁殖生长状况.In order to forecast the occurrence of red tide in the ocean, it is necessary to observe the state of algae reproduction in seawater. A model is proposed to predict the concentration of chlorophyll-a for forecasting the state of algae reproduction indirectly. A data driven learning algorithm of fuzzy nearest neighbor clustering is used for the pretreatment of sampling data. The state prediction model for concentration of chlorophyll-a in seawater is built based on the optimal fuzzy logic system. Considering the character of sampling data and the demand of decreasing the rank of the model, a kind of improved learning algorithm of fuzzy nearest neighbor clustering was designed, and it was used in the state prediction for concentration of chlorophyll-a. The experiments indicate that the improved method can effectively extract the character of sampling data, reduce the dimension of the data, and improve the convergence speed of the learning algorithm. The improved fuzzy nearest neighbor clustering algorithm can reduce the complexity of modeling for prediction, and this kind of mod- el has fine performance of generalization. It can effectively predict the state of algae reproduction in seawater.

关 键 词：模糊最近邻聚类 藻类繁殖 叶绿素A 状态预测 泛化性能 

分 类 号：X55[环境科学与工程—环境工程]