基于小波分解的凡纳滨对虾养殖水体水质的仿真研究

Phantom study on water quality dynamic in Litopenaeus vannameiis' s ponds based on wavelet analysis

依据凡纳滨对虾养殖水体中测定的水质数据,利用ARMA模型和神经网络模型两种方法对水质动态进行预测和分析,提出了一种基于小波分解且针对凡纳滨对虾养殖水体水质预测的ARMA模型,且ARMA(p,q)模型中的p值和q值分别为4和2。预测结果表明,所建立的预测模型精度较高。将ARMA模型预测的结果与神经网络预测的结果进行了对比后发现,基于小波分解的ARMA模型对对虾养殖水体水质预测的有效性和准确性优于神经网络预测模型。

Dynamics of water quality was predicted by ARMA Models and Neural Network Models in this paper. A ARMA Model for water quality of Litopenaeus vannameiis＇s ponds has been raised based on data from shrimps ponds, and the values of p and q of ARMA （p, q） were 4 and 2 respectively. Results of prediction indicated that accuracy of ARMA Model was high. Moreover, comparison of effectivity and accuracy between ARMA Model and Neural Network Models indicated that the former was better for prediction of water quality of shrimp ponds.