Predicting Urban Water Consumption and Health Using Artificial Intelligence Techniques in Tanganyika Lake, East Africa

Water, Год журнала: 2024, Номер 16(13), С. 1793 - 1793

Опубликована: Июнь 25, 2024

Water quality has significantly declined over the past few decades due to high industrial rates, rapid urbanization, anthropogenic activities, and inappropriate rubbish disposal in Lake Tanganyika. Consequently, forecasting water quantity is crucial for ensuring sustainable resource management, which supports agricultural, industrial, domestic needs while safeguarding ecosystems. The models were assessed using important statistical variables, a dataset comprising six relevant parameters, use records. database contained electrical conductivity, pH, dissolved oxygen, nitrate, phosphates, suspended solids, temperature, consumption records, an appropriate date. Furthermore, Random Forest, K-nearest Neighbor, Support Vector Machine are three machine learning methodologies employed categorization forecasting. Three recurrent neural networks, namely long short-term memory, bidirectional gated unit, have been specifically designed predict urban index. classification produced by Forest forecast had highest accuracy of 99.89%. GRU model fared better than LSTM BiLSTM with values R2 NSE, 0.81 0.720 0.78 0.759 index, prediction results. outcomes showed how reliable was classifying forecasts units predicting indices demand. It worth noting that accurate predictions essential public health protection, ecological preservation. Such promising research could enhance demand planning management.

Язык: Английский

---

Recent Progress on Surface Water Quality Models Utilizing Machine Learning Techniques

Water, Год журнала: 2024, Номер 16(24), С. 3616 - 3616

Опубликована: Дек. 15, 2024

Surface waterbodies are heavily exposed to pollutants caused by natural disasters and human activities. Empowering sensor technologies in water quality monitoring, sufficient measurements have become available develop machine learning (ML) models. Numerous ML models quickly been adopted predict indicators various surface waterbodies. This paper reviews 78 recent articles from 2022 October 2024, categorizing utilizing into three groups: Point-to-Point (P2P), which estimates the current target value based on other at same time point; Sequence-to-Point (S2P), utilizes previous series data one point ahead; Sequence-to-Sequence (S2S), uses forecast sequential values future. The used each group classified compared according indicators, availability, model performance. Widely strategies for improving performance, including feature engineering, hyperparameter tuning, transfer learning, recognized described enhance effectiveness. interpretability limitations of applications discussed. review provides a perspective emerging

Язык: Английский

---

Common issues of data science on the eco-environmental risks of emerging contaminants

Environment International, Год журнала: 2025, Номер 196, С. 109301 - 109301

Опубликована: Янв. 27, 2025

Язык: Английский

---

An Analysis of Quality Parameters Changes in Agricultural Water Systems with Wavelet Transform Model

Water, Год журнала: 2025, Номер 17(5), С. 662 - 662

Опубликована: Фев. 25, 2025

Climate change, population growth, industrialization, overconsumption, and pollution strain water resources, posing risks to ecosystem sustainability. Urgent action plans based on decision support systems are essential protect environmental health secure global food resources. This study employs the Wavelet model analyze impacts of agricultural factors resources in a selected irrigation basin by assessing quality parameters, including chemical, physical, biological properties, through seasonal sampling wavelet transformations detect temporal spatial trends. Results showed increased salinity, nitrate, boron, total suspended solids (TSS), chemical oxygen demand (COD) groundwater canals, particularly during dry periods. High nitrate (average 0.36 mg/L) TSS levels 1152 were linked activities, while industrial influences contributed variability boron ranging from 0.01 0.40 mg/L COD 20 235 mg/L. The highlights persistence challenges differences driven external factors. Predictive analyses suggest that without intervention, could worsen. These findings highlight need for wavelet-based techniques develop accurate management strategies mitigating ensuring long-term resource sustainability irrigation-dependent regions.

Язык: Английский

---

Metamodeling of a Physically Based Pesticide Runoff Model with a Long Short-Term Memory Approach

Опубликована: Янв. 1, 2025

Язык: Английский

---

Long Short-Term Memory (LSTM) Networks for Accurate River Flow Forecasting: A Case Study on the Morava River Basin (Serbia)

Water, Год журнала: 2025, Номер 17(6), С. 907 - 907

Опубликована: Март 20, 2025

Accurate forecasting of river flows is essential for effective water resource management, flood risk reduction and environmental protection. The ongoing effects climate change, in particular the shift precipitation patterns increasing frequency extreme weather events, necessitate development advanced models. This study investigates application long short-term memory (LSTM) neural networks predicting runoff Velika Morava catchment Serbia, representing a pioneering LSTM this region. uses daily runoff, temperature data from 1961 to 2020, interpolated using inverse distance weighting method. model, which was optimized trial-and-error approach, showed high prediction accuracy. For station, model mean square error (MSE) 2936.55 an R2 0.85 test phase. findings highlight effectiveness capturing nonlinear hydrological dynamics, temporal dependencies regional variations. underlines potential models improve management strategies Western Balkans.

Язык: Английский

---

Impact of Seasonal Variation and Population Growth on Coliform Bacteria Concentrations in the Brunei River: A Temporal Analysis with Future Projection

Water, Год журнала: 2025, Номер 17(7), С. 1069 - 1069

Опубликована: Апрель 3, 2025

Coliform bacteria pollution poses a significant challenge to water quality in the Brunei River, critical resource Darussalam. This study investigates impact of seasonal variations and population growth on coliform concentrations across eight monitoring stations while addressing data limitations forecasting future trends. Seasonal variations, analyzed using box plots, revealed significantly higher levels during rainy season, driven by urban residential runoff. Population growth, assessed propensity score matching, showed that densely populated areas experienced elevated contamination levels. Temporal trends, Rescaled Adjusted Partial Sums (RAPS) method, indicated declining trend from 2013 2018, followed sharp increase post-2018, linked urbanization, wastewater discharge, overburdened sewage infrastructure, particularly upstream stations. To forecast levels, ARIMA, Logistic Regression, Bidirectional Long Short-Term Memory (BiLSTM) models were employed their predictive performance evaluated. Despite constraints small dataset, BiLSTM model outperformed others most stations, emphasizing its ability capture complex temporal relationships. Furthermore, Mann–Kendall analysis predicted over five-year period upward trends highlights potential combining advanced with robust analytical techniques focused collection efforts support sustainable management data-scarce environments.

Язык: Английский

---

Probabilistic machine learning-based phytoplankton abundance using hyperspectral remote sensing

GIScience & Remote Sensing, Год журнала: 2025, Номер 62(1)

Опубликована: Апрель 11, 2025

Язык: Английский

---

Non-Linear Synthetic Time Series Generation for Electroencephalogram Data Using Long Short-Term Memory Models

AI, Год журнала: 2025, Номер 6(5), С. 89 - 89

Опубликована: Апрель 25, 2025

Background/Objectives: The implementation of artificial intelligence-based systems for disease detection using biomedical signals is challenging due to the limited availability training data. This paper deals with generation synthetic EEG deep learning-based models, be used in future research Parkinson’s systems. Methods: Linear such as AR, MA, and ARMA, are often inadequate inherent non-linearity time series. To overcome this drawback, long short-term memory (LSTM) networks proposed learn long-term dependencies non-linear series subsequently generate enhance forward backward signals, a Bidirectional LSTM model has been implemented. was trained on UC San Diego Resting State Dataset, which includes samples from two groups: individuals healthy control group. Results: determine optimal number cells model, we evaluated mean squared error (MSE) cross-correlation between original signals. method also applied select length hidden state vector. set 14, vector each cell fixed at 4. Increasing these values did not improve MSE or unnecessarily increased computational complexity. model’s performance mean-squared (MSE), Pearson’s correlation coefficient, power spectra demonstrating suitability application. Conclusions: compared Autoregressive Moving Average (ARMA) superior performance. confirms that LSTM, strong alternatives statistical models like ARMA handling non-linear, multifrequency, non-stationary

Язык: Английский

---

A Mamba-based method for multi-feature water quality prediction fusing dual denoising and attention enhancement

Journal of Hydrology, Год журнала: 2025, Номер unknown, С. 133424 - 133424

Опубликована: Апрель 1, 2025

Язык: Английский

---