基于BP网络养殖水体氨氮预测模型及实现  

Prediction Model Based on Neural Network for NH_4^+-N and Realization

作　　者：高艳萍[1] 周敏[1] 姜凤娇[1] 

机构地区：[1]大连水产学院信息工程学院,辽宁大连116023

出　　处：《农机化研究》2008年第7期48-50,共3页Journal of Agricultural Mechanization Research

基　　金：辽宁省教育厅攻关项目(05L090)

摘　　要：在分析水产养殖水质影响因素的基础上,利用BP神经网络良好的非线性映射特性,建立BP网络预测模型预测水产养殖水体氨氮的含量,采用最速梯度下降法优化全值对预测模型进行训练。训练结果表明:预测模型收敛速度快,预测精度高,对水产养殖智能控制的实现提供了可行方案。On the basis of analyzing parameter about fishery water, modeling method of prediction for NH4^＋ -N of fishery water was designed using better nonlinear approaching ability of neural network . Value of weight was optimized with dropping of differential of error rapidly, and training it, The results of training shows： it is convergent quickly and prediction result is more precise too. It present an efficient scheme for realizing artificial control of fishery culture.

关 键 词：神经网络 MATLAB 预测模型 氨氮 优化算法 

分 类 号：S967.4[农业科学—水产养殖] TP183[农业科学—水产科学]