Prediction of Inland Aquaculture Ammonia Using Hybrid Intelligent Soft Computing

One of the crucial factors in assessing the pond's intensive inland aquaculture water quality condition is ammonia. The excessive ammonia content will likely worsen water quality and result in the mass mortality of cultured individuals. For aquaculture management, it is therefore vital to accurately identify the ammonia nitrogen level of cultured water. However, the accuracy of technology for monitoring the ammonia content of aquaculture water currently needs to be improved to satisfy the demands of intensive aquaculture. This paper presents the prediction of the ammonia concentration of aquaculture water in real time using a hybrid intelligent soft computing algorithm. Radial basis function neural networks (RBFNN) and a hybrid model combining RBFNN, and particle swarm optimization (PSO) are used in this technique. Root mean square error (RMSE) and correlation coefficient (R²) were two separate statistical metrics used to compare the two methodologies and assess how well the soft computing strategies performed. The ammonia prediction results showed that the PSO-RBFNN method outperformed the RBFNN. The PSO-RBFNN model offers a real-time ammonia prediction value in inland farming waters that is moderately and generally accurate.