PREDICTION OF AMMONIACAL NITROGEN IN RIVER USING MODIFIED CUCKOO SEARCH ľNEURAL NETWORK (CS-NN)Siti Fatimah Che Osmi1, M.A. Malek and M. Yusoff
The avant-garde of Artificial Neural Network (ANN) for water quality prediction provides new interest to researchers, experts, and practitioners in environmental engineering to explore and improve the ability of ANN using coupled/hybrid optimization method to reduce error and increase efficiency. Indeed, ANN has many advantages in term of data utilization, knowledge practice, cost, and time consumed which can further improve current deterministic model. In this study, Modified ANN with Cuckoo-Search (CS) is proposed to improve water quality monitoring and surveillance at the study area. The results demonstrated the ability of CS in improving BPNN models for prediction Ammoniacal Nitrogen (NH3-N) at all river stations with R2 value obtained at more than 0.99 and zero error for Mean Absolute Error (MAE). In conclusion, the proposed Modified CS-NN prediction model is indeed pertinent in enhancing the performance of conventional BPNN model and deterministic models used as an engineering tool for water quality prediction.
Enter your contact information below to receive full paper.