Prediction model of dissolved oxygen in marine pasture based on hybrid gray wolf algorithm optimized support vector regression

Water quality prediction plays a vital role in water pollution warning and control. However, traditional prediction models usually suffer from low efficiency and poor robustness. To predict accurately the dissolved oxygen concentration in the marine pasture, a dissolved oxygen prediction model, based on wavelet analysis and hybrid gray wolf algorithm (HGWO) optimized vector regression, was established. Because of the water quality data in the marine pasture is a stationary time series. To improve the accuracy of water quality data, wavelet analysis was applied for data pre-processing in this paper. Besides, after the gray wolf algorithm (GWO) was optimized by the differential evolution algorithm (DE), it was used to optimize the support vector regression (SVR). Hence, the SVR’s disadvantages of optimization ability and prediction accuracy both were improved. Back propagation neural network (BPNN), SVR, GWO-SVR, DE-SVR, and this model were, respectively, used to predict the dissolved oxygen concentration of Beidaihe marine pasture. The experimental results show that the mean square error, mean absolute error, and average percentage error of the model are 0.1658, 0.359, and 0.0305, respectively, which are better than the traditional prediction model. So this model has higher prediction accuracy and stronger generalization ability, and it can provide a reference for the precise regulation of aquaculture.