In recent years, researchers and policymakers have focused on leaks in water systems as a critical issue because of their negative impact on human society. Most classical methods can only provide approximate leakage locations, typically identifying the general area of a node or pipe section in the network. In this paper, frequency ratio (FR) and deep learning neural network (DLNN) models were used to identify water leaks in water distribution networks (WDNs). Eight predictor variables were used to assess leakage susceptibility in the WDN of Khenchela City, in northeastern Algeria. According to both models, the two most significant predictor variables in the studied WDN are pipe material and age. The predictive ability of these models was evaluated using the receiver operating characteristic (ROC) curve, based on 339 recorded water leakage locations. The results were very satisfactory, with the DLNN model showing slightly higher accuracy than the FR model, achieving area under the curve (AUC) values of 86.7% and 82.3%, respectively. Although the FR model was applied for the first time in this field, it demonstrated strong potential as a decision-support tool for water leak detection.

Performance comparison of tree-based and neural network models for wear prediction in coated and uncoated Al6061

Exploring aroma descriptions of different cherry juice and the mechanism of aroma formation in Lapins using volatilomics and machine learning.

Imidazole functionalized nanozyme for deep detoxification and machine-learning-assisted intelligent sensing of profenofos.

Industrial Internet of Things intrusion detection based on a hybrid model of Pearson-Deep Neural Network And Transformer

Advancing post-combustion CO2 capture: experimental and theoretical analysis of potassium sarcosine/MDEA solutions via artificial neural network, Deshmukh-Mather, and semi-empirical models

Retraction notice to “A robust solution for recognizing accurate handwritten text extraction using quantum convolutional neural network and transformer models” [Computers and Electrical Engineering 120 (2024) 109794

Deep learning-based ROSC prediction and ECG phenotyping in out-of-hospital cardiac arrest.

Optimization of blue laser welding for copper hairpins in EV motors

Image-based prediction of soiling-induced power loss in solar panels: A novel neural architecture search method via reinforcement learning

Explainable deep learning for grading of Edible Bird's Nest (EBN)

The recovery of valuable metals from e-waste leachates is essential for advancing circular economy strategies and reducing environmental risks. This study examined Low Density Concrete (LDC), a waste material, as a sustainable adsorbent for the recovery of Ni 2+ , Mn 2+ , Zn 2+ , and Fe 2+ /Fe 3+ . This is the first study that evaluated the performance of LDC in metal recovery from a real leachate produced by the anaerobic digestion of alkaline batteries and municipal solid waste. The one-factorial method was employed to find the effect of the pH and adsorbent mass on results. The adsorption behavior was then studied using three different isothermal models: Freundlich, Langmuir, and Temkin. The findings indicated that pH significantly influenced metal removal, with ion exchange predominating in acidic conditions (pH < 5) and adsorption-precipitation mechanisms becoming more significant near neutral pH. Optimal performance was achieved at pH = 7 and an adsorbent dosage of 0.10 g. The most significant parameter influencing metal removal efficiency was pH, as determined by ANOVA. To enhance process prediction, both Response Surface Methodology (RSM) and Artificial Neural Network (ANN) models were developed. ANN had a better predictive accuracy ( r > 0.98) than RSM. Material characterization such as FTIR, SEM-EDS, and TGA, confirmed metal uptake and associated surface and structural changes. Finally, an environmental impact assessment using the Leopold Matrix indicated that LDC exhibits lower environmental impacts compared to conventional adsorbents. These findings support the potential of LDC as a green, low-cost material for metal recovery from complex e-waste leachates. • Recovered Ni 2+ , Mn 2+ , Zn 2+ , Fe 2+ /Fe 3+ ions from e-waste leachate using concrete waste. • LDC showed dual adsorption–ion exchange behavior under varying pH conditions. • ANN predicted adsorption capacity with r > 0.98, surpassing RSM performance. • Freundlich model confirmed multilayer adsorption on heterogeneous LDC surface. • EIA proved LDC greener than GO regarding water, energy, and human health impact.

An online SO2 and CS2 detection system combining weighted elliptical dual-spectrum reconstruction (WEDSR) with convolutional neural network (CNN).

Diabetic foot disorders continue to be among the most prevalent and overlooked complications associated with diabetes. The aim of this study was to determine the factors associated with diabetic foot complications in semi-urban Udupi District. The study was a cross-sectional study. 25,000 individuals living in Udupi district were screened for diabetes mellitus, and among them, 3844 individuals were found to have type-2 diabetes mellitus (T2DM). Further, detailed anthropometry and foot assessments were performed for these individuals. In this study, a total of 3844 participants aged between 40 and 75 years with T2DM were screened to determine the prevalence of foot complications. The mean age of the study participants was 59.2 years (±11.7). Of the participants, 41.3% were male and 58.7% were female. Neuropathy was present in 9.8% of the participants, and 5.6% of the participants had a foot ulcer. Among 3844 individuals, sensation, pedal pulse, vibration, and foot care awareness were factors associated with diabetic foot complications. The Bayesian Neural Network (BNN) model was also developed, and showed good predictive performance, with an AUC of 0.901 for the right foot and 0.922 for the left foot. The BNN results also show strong predictive performance. Both models predicted diabetic foot complications. Prevalence of foot complications is high in the Udupi district, and the presence of risk factors puts the individual at risk for serious complications of T2DM.

Extreme flooding poses escalating risks to roadway infrastructure, threatening structural integrity, operational reliability, and public safety. Existing flood-monitoring approaches primarily utilize sparse sensor networks, thus providing limited real-time data and insufficiently capturing indirect flooding impacts on noninundated roadway segments. This gap complicates emergency response, delays evacuation, and undermines postevent recovery efforts. Addressing this critical knowledge gap, this study proposes an advanced predictive decision-support framework leveraging an artificial intelligence (AI)-enhanced digital twin integrated with flood simulations and graph neural network (GNN) modeling. It systematically assesses roadway serviceability during extreme flooding by integrating structural conditions, operational disruptions, historical maintenance records, inundation severity, and recovery timelines. By consolidating heterogeneous data sets, including historical traffic volumes, pavement conditions, hydrological data, and weather forecasts, the developed framework provides accurate, real-time predictive insights for both inundated and indirectly impacted roadway segments. This capability was demonstrated through an illustrative validation study. Ultimately, this research equips transportation agencies and emergency responders with robust hands-on tools that can facilitate optimized emergency response, improved infrastructure management, and enhanced resilience of transportation systems under ever-increasing flood risks due to climate change.

Automatic classification of orange fruit diseases using deep neural network model

Quantile-based interpretable neural network models: Mortality forecasting and actuarial simulations

Ternary heterojunction Bi2WO6/TiO2/GO photocatalyst for the removal of Cefixime: Bridging experimental characterizations, photocatalytic reaction-informed neural network modeling, genetic algorithm optimization, and density functional theory computations

A new artificial neural network model for prediction of fatigue strength and yield strength of various steel grades

Two-phase collaborative model compression training for joint pruning and quantization.