Study on Prediction of Dissolved Oxygen Content in Aquaculture Water

Abstract

Aiming at the problems of low accuracy, slow convergence and poor robustness of traditional neural network water quality prediction method, a dissolved oxygen content prediction model based on combining algorithm of improved Fruit fly optimization algorithm and BP neural network (IFOABP) is proposed. The best combination of weights and biases parameters of BP neural network is obtained by improved Fruit fly optimization algorithm, and the prediction model of dissolved oxygen content in water quality is established. The model is applied to the prediction and analysis of dissolved oxygen in Zhangjialou Breeding Base in Qingdao. The experimental results show that the model has better prediction effect than BP neural network, FOA-BP and GA-BP. The mean absolute percentage error (MAPE), root mean square error (RMSE), mean absolute error (MAE) and determination coefficient (R2) of IFOA-BP are 0.4013 and 0.1346, 0.0626, 0.9989. The BP neural network optimized in this paper not only has fast convergence speed and high prediction accuracy, but also provides a reliable decision basis for dissolved oxygen control in intensive aquaculture water.