The reliable prediction of dissolved oxygen concentration (DO) is significantly crucial for protecting the health of the aquatic ecosystem. The current research employed four different single AI-based models, namely long short-term memory neural network (LSTM), extreme learning machine (ELM), Hammerstein-Weiner (HW) and general regression neural network (GRNN) for modeling the DO concentration of Kinta River, Malaysia using available water quality (WQ) parameters. Afterwards, the first scenario used four different ensemble techniques (ET). Two linear, i.e. simple averaging ensemble (SAE) and weighted averaging ensemble (WAE) and two nonlinear namely; backpropagation neural network ensemble (BPNN-E) and HW ensemble (HW-E). The second scenario employed a hybrid random forest (RF) ensemble in order to enhance the prediction accuracy of the single models. The WQ parameters were subjected to a different pre-analysis test to ascertain their stability. The four-model combinations are generated using the nonlinear sensitivity input selection approach. The modeling performance was assessed using the statistical measures of Nash-Sutcliffe coefficient efficiency (NSE), Willmott’s index of agreement (WI), root mean square error (RMSE), mean absolute error (MAE) and mean square error (MSE) and correlation coefficient (CC). The results of the single AI-based models demonstrated that HW (M3) served as the best model for predicting DO concentration. For ensemble results, BPNN-E (WI=0.9764) was superior to the other three ET with average decreased of more than 2% with regards to MAE. Investigation on the hybrid RF ensemble demonstrated the reliable accuracy for all the hybrid models with better predictive skill shown by the HW-RF (CC=0.981) ensemble. The overall results verified the promising impact of HW-M3, ET and hybrid RF ensemble for the prediction of the DO concentration in the Kinta River, Malaysia.

Modelling of a river involves protracted engagement with uncertainty, thus making developing a reliable model becomes quite cumbersome and often impossible. This paper presents a comparison of a non-linear models based on Radial Basis Function Neural Network (RBFNN) and Hammerstein-Weiner Model (HW) and a conventional linear models, Auto regressive Integrated Moving Average (ARIMA) and Generalized Linear Regression (GLR) models for Kinta River, in Malaysia and Agra station of Yamuna river in India. Experimental data from the rivers were used in validating the models. Simulation results demonstrated that the nonlinear models (RBFNN and HW) averagely increased the performance accuracy of the linear models (ARIMA and GLR) by 20% and 15% at Kinta and Agra station of Yamuna River in the verification phase respectively. Considering the overall results the RBFNN model outperformed the other models with the average increase up to 21% in the verification phase. The model could serves as reliable and useful tool for forecasting the rivers.

Procurement of measuring device, maintenance cost coupled with calibration of the instrument contributed to the difficulty in forecasting of global solar radiation in underdeveloped countries. Most of the available regressional and mathematical models do not capture well the behavior of the global solar radiation. This paper presents the comparison of Adaptive Neuro Fuzzy Inference System (ANFIS) and Autoregressive Moving Average with eXogenous term (ARMAX) in forecasting global solar radiation. Full-Scale (experimental) data of Nigerian metrological agency, Sultan Abubakar III international airport Sokoto was used to validate the models. The simulation results demonstrated that the ANFIS model having achieved MAPE of 5.34% outperformed the ARMAX model. The ANFIS could be a valuable tool for forecasting the global solar radiation.

A dependable design and monitoring of solar energy-based systems necessitates precise data on available solar radiation. However, measuring solar radiation is challenging due to the expensive equipment required for measurement, along with the costs of calibration and maintenance, especially in developing countries like Nigeria. As a result, data-driven techniques are often employed to predict solar radiation in such regions. However, the existing predictive models frequently yield unsatisfactory outcomes. To address this issue, this study proposes the creation of intelligent models to forecast solar radiation in Kano state, Nigeria. The model is developed using an ensemble machine learning approach that combines two Adaptive Neuro-Fuzzy Inference Systems with sub-clustering optimization and grid-partitioning optimization. The meteorological data used for model development include maximum temperature, minimum temperature, mean temperature, and solar radiation from the previous 2 days as predictors. To evaluate the models performance, various metrics like correlation coefficient, determination coefficient, mean-squared error, root-mean-squared error, and mean-absolute error are employed. The simulation results demonstrate that the ANFIS ensemble outperforms the individual ANFIS models. Notably, the ANFIS-ENS exhibits the highest accuracy. Consequently, the developed models provide a reliable alternative for estimating solar radiation in Kano and can be instrumental in enhancing the design and management of solar energy systems in the region.