Functionalization of remote sensing and on-site data for simulating surface water dissolved oxygen: Development of hybrid tree-based artificial intelligence models.

Dissolved oxygen (DO) is an important indicator of river health for environmental engineers and ecological scientists to understand the state of river health. This study aims to evaluate the reliability of four feature selector algorithms i.e., Boruta, genetic algorithm (GA), multivariate adaptive regression splines (MARS), and extreme gradient boosting (XGBoost) to select the best suited predictor of the applied water quality (WQ) parameters; and compare four tree-based predictive models, namely, random forest (RF), conditional random forests (cForest), RANdom forest GEneRator (Ranger), and XGBoost to predict the changes of dissolved oxygen (DO) in the Klang River, Malaysia. The total features including 15 WQ parameters from monitoring site data and 7 hydrological components from remote sensing data. All predictive models performed well as per the features selected by the algorithms XGBoost and MARS in terms applied statistical evaluators. Besides, the best performance noted in case of XGBoost predictive model among all applied predictive models when the feature selected by MARS and XGBoost algorithms, with the coefficient of determination (R2) values of 0.84 and 0.85, respectively, nonetheless the marginal performance came up by Boruta-XGBoost model on in this scenario.

Prediction of lead (Pb) adsorption on attapulgite clay using the feasibility of data intelligence models.

This study investigates the performance of support vector machine (SVM), multivariate adaptive regression spline (MARS), and random forest (RF) models for predicting the lead (Pb) adsorption by attapulgite clay. Models are constructed using batch stochastic data of heavy metal (HM) concentrations under different physicochemical conditions. Implementation of auto-hyper-parameter tuning using grid-search approach and comparative analysis is performed against the benchmark artificial intelligence (AI) models. Models are constructed based on Pb concentration (IC), the dosage of attapulgite clay (dose), contact time (CT), pH, and NaNO3 (SN). Principle component analysis (PCA) and correlation analysis (CA) methods are integrated to assess the importance of the applied predictors and their relationship with the target. Research findings approved the potential of the grid-RF model as a marginal superior predictive model against the grid-SVM in terms of MAE, i.e., 3.29 and 3.34, respectively; moreover, the md scored the same, i.e., 0.93, which reveals the potential predictability for both. Nonetheless, grid-MARS and standalone MARS models remained likewise in their predictability. IC parameter demonstrated the highest influential among all the predictors with the highest value of importance in the case of all three evaluators. The solution pH and dose stands together with marginal differences in case of PCA method; however, solution pH and CT appeared with similarity impact using the PCA method.

Prediction of sediment heavy metal at the Australian Bays using newly developed hybrid artificial intelligence models.

Hybrid artificial intelligence (AI) models are developed for sediment lead (Pb) prediction in two Bays (i.e., Bramble (BB) and Deception (DB)) stations, Australia. A feature selection (FS) algorithm called extreme gradient boosting (XGBoost) is proposed to abstract the correlated input parameters for the Pb prediction and validated against principal component of analysis (PCA), recursive feature elimination (RFE), and the genetic algorithm (GA). XGBoost model is applied using a grid search strategy (Grid-XGBoost) for predicting Pb and validated against the commonly used AI models, artificial neural network (ANN) and support vector machine (SVM). The input parameter selection approaches redimensioned the 21 parameters into 9-5 parameters without losing their learned information over the models' training phase. At the BB station, the mean absolute percentage error (MAPE) values (0.06, 0.32, 0.34, and 0.33) were achieved for the XGBoost-SVM, XGBoost-ANN, XGBoost-Grid-XGBoost, and Grid-XGBoost models, respectively. At the DB station, the lowest MAPE values, 0.25 and 0.24, were attained for the XGBoost-Grid-XGBoost and Grid-XGBoost models, respectively. Overall, the proposed hybrid AI models provided a reliable and robust computer aid technology for sediment Pb prediction that contribute to the best knowledge of environmental pollution monitoring and assessment.

Heavy metal contamination prediction using ensemble model: Case study of Bay sedimentation, Australia.

Lead (Pb) is a primary toxic heavy metal (HM) which present throughout the entire ecosystem. Some commonly observed challenges in HM (Pb) prediction using artificial intelligence (AI) models include overfitting, normalization, validation against classical AI models, and lack in learning/technology transfer. This study explores the extreme gradient boosting (XGBoost) model as a superior SuperLearning (SL) algorithms for Pb prediction. The proposed model was examined using historical data at the Bramble and Deception Bay (BB and DB) stations, Australia. The model was trained at one of the stations and transferred to a cross-station and vice versa. XGBoost showed higher reliability with less declination in (R2: coefficient of determination), i.e., 0.97 % over the testing phase, among others models at BB. At the cross-station (DB), the performance of the XGBoost model was decreased by 2.74 % (R2) against random forests (RF). The mean absolute error (MAE) observed 40 % (XGBoost) and 47 % (RF) less than artificial neural network (ANN). The XGBoost model performance declined by 3.44 % (R2) over testing (DB), which is minor among validated models. At the cross-station (BB), the XGBoost model showed the least decrement in terms of R2, i.e., 7.99 % against the ANN (8.31 %), RF (10.26 %), and support vector machine (SVM, 36.19 %).

Manganese (Mn) removal prediction using extreme gradient model.

Exploring the Manganese (Mn) removal prediction with several independent variables is tremendously critical and indispensable to understand the pattern of removal process. Mn is one of the key heavy metals (HMs) stipulated by the WHO for the development of many attributes of the ecosystem in controlled quantity. In the present paper, an extreme gradient model (XGBoost) is proposed for Mn prediction. A compressive statistical analysis reveals the stochastics behaviour of the data prior to the prediction investigation. The main goal is to determine the Mn predictability of XGBoost algorithm with influencing factors such as D2EHPA (M), Time (min), H2SO4 (M), NaCl (g/L), and EDTA (mM). The PCA biplot signifies the importance of the predictors. The XGBoost model validated against a diversity of data-driven models such as multilinear regression (MLR), support vector machine (SVM), and random forest (RF). The order of the applied models' performance are XGBoost > RF > SVM > MLR as per their R2 and RMSE metrics over testing phase i.e. 20.88, 0.75, 0.61, 0.40, and 2.23, 3.01, 3.51, 6.38, respectively. Moreover, the Taylor diagram and Radar chart have drown to emphasize the XGBoost model efficiency, stability, and reliability. In respect of XGBoost model prediction, 'Time' predictor outperforms D2EHPA, EDTA, H2SO4, and NaCl predictors in order.

Laundry wastewater treatment using a combination of sand filter, bio-char and teff straw media.

Numerous researchers have expressed concern over the emerging water scarcity issues around the globe. Economic water scarcity is severe in the developing countries; thus, the use of inexpensive wastewater treatment strategies can help minimize this issue. An abundant amount of laundry wastewater (LWW) is generated daily and various wastewater treatment researches have been performed to achieve suitable techniques. This study addressed this issue by considering the economic perspective of the treatment technique through the selection of easily available materials. The proposed technique is a combination of locally available absorbent materials such as sand, biochar, and teff straw in a media. Biochar was prepared from eucalyptus wood, teff straw was derived from teff stem, and sand was obtained from indigenous crushed stones. In this study, the range of laundry wastewater flow rate was calculated as 6.23-17.58 m3/day; also studied were the efficiency of the media in terms of the removal percentage of contamination and the flux rate. The performances of biochar and teff straw were assessed based on the operation parameters and the percentage removal efficiency at different flux rates; the assessment showed 0.4 L/min flux rate to exhibit the maximum removal efficiency. Chemical oxygen demand, biological oxygen demand, and total alkalinity removal rate varied from 79% to ≥83%; total solids and total suspended solids showed 92% to ≥99% removal efficiency, while dissolved oxygen, total dissolved solids, pH, and electrical conductivity showed 22% to ≥62% removal efficiency. The optimum range of pH was evaluated between 5.8-7.1. The statistical analysis for finding the correlated matrix of laundry wastewater parameters showed the following correlations: COD (r = -0.84), TS (r = -0.83), and BOD (r = -0.81), while DO exhibited highest negative correlation. This study demonstrated the prospective of LWW treatment using inexpensive materials. The proposed treatment process involved low-cost materials and exhibited efficiency in the removal of contaminants; its operation is simple and can be reproduced in different scenarios.