Machine learning model combined with CEEMDAN algorithm for monthly precipitation prediction

Earth Science Informatics, Journal Year: 2023, Volume and Issue: 16(2), P. 1821 - 1833

Published: April 26, 2023

Language: Английский

---

Prediction of air quality index based on the SSA-BiLSTM-LightGBM model

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: April 5, 2023

The air quality index (AQI), as an indicator to describe the degree of pollution and its impact on health, plays important role in improving atmospheric environment. Accurate prediction AQI can effectively serve people's lives, reduce control costs improve In this paper, we constructed a combined model based real hourly data Beijing. First, used singular spectrum analysis (SSA) decompose into different sequences, such trend, oscillation component noise. Then, bidirectional long short-term memory (BiLSTM) was introduced predict decomposed data, light gradient boosting machine (LightGBM) integrate predicted results. experimental results show that effect SSA-BiLSTM-LightGBM for set is good test set. root mean squared error (RMSE) reaches 0.6897, absolute (MAE) 0.4718, symmetric percentage (SMAPE) 1.2712%, adjusted R2 0.9995.

Language: Английский

---

Performance of Bayesian Model Averaging (BMA) for Short-Term Prediction of PM10 Concentration in the Peninsular Malaysia

Atmosphere, Journal Year: 2023, Volume and Issue: 14(2), P. 311 - 311

Published: Feb. 4, 2023

In preparation for the Fourth Industrial Revolution (IR 4.0) in Malaysia, government envisions a path to environmental sustainability and an improvement air quality. Air quality measurements were initiated different backgrounds including urban, suburban, industrial rural detect any significant changes parameters. Due dynamic nature of weather, geographical location anthropogenic sources, many uncertainties must be considered when dealing with pollution data. recent years, Bayesian approach fitting statistical models has gained more popularity due its alternative modelling strategy that accounted all Therefore, this study aims evaluate performance Model Averaging (BMA) predicting next-day PM10 concentration Peninsular Malaysia. A case utilized seventeen years’ worth monitoring data from nine (9) stations located using eight parameters, i.e., PM10, NO2, SO2, CO, O3, temperature, relative humidity wind speed. The performances prediction calculated five models’ evaluators, namely Coefficient Determination (R2), Index Agreement (IA), Kling-Gupta efficiency (KGE), Mean Absolute Error (MAE), Root Squared (RMSE) Percentage (MAPE). BMA indicate humidity, speed contributed most model majority (R2 = 0.752 at Pasir Gudang station), 0.749 Larkin 0.703 Kota Bharu 0.696 Kangar station) 0.692 Jerantut respectively. Furthermore, demonstrated good performance, IA ranging 0.84 0.91, R2 0.64 0.75 KGE 0.61 0.74 stations. According results investigation, should utilised research forecasting operations pertaining issues such as pollution. From study, is recommended one tools concentration, especially particulate matter level.

Language: Английский

---

Co-Training Semi-Supervised Learning for Fine-Grained Air Quality Analysis

Atmosphere, Journal Year: 2023, Volume and Issue: 14(1), P. 143 - 143

Published: Jan. 9, 2023

Due to the limited number of air quality monitoring stations, data collected are limited. Using supervised learning for fine-grained analysis, that is used predict index (AQI) locations without may lead overfitting in models have superior performance on training set but perform poorly validation and testing set. In order avoid this problem learning, most effective solution increase amount data, study, not realistic. Fortunately, semi-supervised can obtain knowledge from unlabeled samples, thus solving caused by insufficient samples. Therefore, a co-training method combining K-nearest neighbors (KNN) algorithm deep neural network (DNN) proposed, named KNN-DNN, which makes full use samples improve model analysis. Temperature, humidity, concentrations pollutants source type as input variables, KNN DNN learners. For each learner, labeled initial relationship between variables AQI. iterative process, labeling pseudo-sample with highest confidence selected expand The proposed evaluated real dataset stations 1 February 30 April 2018 over region 118° E–118°53′ E 39°45′ N–39°89′ N. Practical application shows has significant effect analysis quality. coefficient determination predicted value true 0.97, better than other models.

Language: Английский

---

Natural gas demand response strategy considering user satisfaction and load volatility under dynamic pricing

Energy, Journal Year: 2023, Volume and Issue: 277, P. 127725 - 127725

Published: May 2, 2023

Language: Английский

---

Predicting air quality index using attention hybrid deep learning and quantum-inspired particle swarm optimization

Journal Of Big Data, Journal Year: 2024, Volume and Issue: 11(1)

Published: May 11, 2024

Abstract Air pollution poses a significant threat to the health of environment and human well-being. The air quality index (AQI) is an important measure that describes degree its impact on health. Therefore, accurate reliable prediction AQI critical but challenging due non-linearity stochastic nature particles. This research aims propose hybrid deep learning model based Attention Convolutional Neural Networks (ACNN), Autoregressive Integrated Moving Average (ARIMA), Quantum Particle Swarm Optimization (QPSO)-enhanced-Long Short-Term Memory (LSTM) XGBoost modelling techniques. Daily data were collected from official Seoul registry for period 2021 2022. first preprocessed through ARIMA capture fit linear part followed by architecture developed in pretraining–finetuning framework non-linear data. used convolution extract features original data, then QPSO optimize hyperparameter LSTM network mining long-terms time series features, was adopted fine-tune final model. robustness reliability resulting assessed compared with other widely models across meteorological stations. Our proposed achieves up 31.13% reduction MSE, 19.03% MAE 2% improvement R-squared best appropriate conventional model, indicating much stronger magnitude relationships between predicted actual values. overall results show attentive inspired more feasible efficient predicting at both city-wide station-specific levels.

Language: Английский

---

An All-Hazards Return on Investment (ROI) Model to Evaluate U.S. Army Installation Resilient Strategies

Systems, Journal Year: 2025, Volume and Issue: 13(2), P. 90 - 90

Published: Jan. 31, 2025

The paper describes our project to develop, verify, and deploy an All-Hazards Return of Investment (ROI) model for the U. S. Army Engineer Research Development Center (ERDC) provide army installations with a decision support tool evaluating strategies make existing installation facilities more resilient. need increased resilience extreme weather caused by climate change was required U.S. code DoD guidance, as well strategic plan that stipulated ROI evaluate relevant resilient strategies. During project, ERDC integrated University Arkansas designed into new planning expanded scope options from all hazards. Our methodology included research on policy, data sources, options, analytical techniques, along stakeholder interviews weekly meetings developers. uses standard risk analysis engineering economics terms analyzes potential hazards using in tool. calculates expected net present cost without strategy, each strategy. minimum viable product formulated mathematically, coded Python, verified hazard scenarios, provided implementation.

Language: Английский

---

Prediction of PM_2.5 and CO₂ concentrations using the PCA-LightGBM method in Bandung, Indonesia

Journal of Physics Conference Series, Journal Year: 2025, Volume and Issue: 2942(1), P. 012004 - 012004

Published: Feb. 1, 2025

Abstract Poor air quality due to large amounts of human activity shows the need increase public awareness and alertness by building a system predicting future pollutant concentrations. This research creates prediction using LightGBM algorithm for PM 2.5 CO 2 parameters with an additional parameter reduction method PCA accuracy. The number valid datasets is 918 each five at measurement station, data gaps filled median values so that they can be used predictions. results show best accuracy Deli which uses MAPE 21.5%, , it achieved station without 4.8%. Based on its accuracy, less suitable if there are outliers in dataset, but ideal homogeneous datasets. Overall, based feasible category, accurate very category. To optimize results, especially long term, necessary retrain complete up-to-date dataset better suit conditions.

Language: Английский

---

Exploring Natural and Anthropogenic Drivers of PM2.5 Concentrations Based on Random Forest Model: Beijing–Tianjin–Hebei Urban Agglomeration, China

Atmosphere, Journal Year: 2023, Volume and Issue: 14(2), P. 381 - 381

Published: Feb. 15, 2023

PM2.5 is the key reason for frequent occurrence of smog; therefore, identifying its driving factors has far-reaching significance prevention and control air pollution. Based on long-term remote sensing inversion data, 21 in fields nature humanities were selected, random forest model was applied to study influencing concentration Beijing–Tianjin–Hebei urban agglomeration (BTH) from 2000 2016. The results indicate: (1) main affecting not only include natural such as sunshine hours (SSH), relative humidity (RHU), elevation (ELE), normalized difference vegetation index (NDVI), wind speed (WIN), average temperature (TEM), daily range (TEMR), precipitation (PRE), but also human urbanization rate (URB), total investment fixed assets (INV), number employees secondary industry (INDU); (2) changed into an inverted S-shape with increase SSH WIN, RHU, NDVI, TEM, PRS, URB INV. As ELE TEMR, it fluctuated decreased ELE, while increased then TEMR. However, change less pronounced PRE INDU; (3) influence higher than that factors, role been continuously strengthened recent years. adjustment pollution sources perspective will become effective way reduce concentrations BTH.

Language: Английский

---

Physicochemical Assessment of the Road Vehicle Traffic Pollution Impact on the Urban Environment

Atmosphere, Journal Year: 2023, Volume and Issue: 14(5), P. 862 - 862

Published: May 11, 2023

Vehicle traffic pollution requires complex physicochemical analysis besides emission level measuring. The current study is focused on two campaigns of emissions measurements held in May and September 2019 Alba Iulia City, Romania. There was found a significant excess PM2.5 for all measuring points PM10 the most circulated during May, along with VOC CO2 emissions. reveal threshold PM increased values These are consequences environmental interaction traffic. Street dust air-suspended particle samples were collected analyzed to evidence sources. Physicochemical investigation reveals highly mineralized particulate matter: fractions within predominantly contain Muscovite, Kaolinite, traces Quartz Calcite, while Calcite. mineral originate street suspended atmosphere due vehicles’ circulation. A amount soot as small micro-sized clusters fine micro-spots attached over particles, observed by Mineralogical Optical Microscopy (MOM) Fourier Transformed Infrared Spectroscopy (FTIR). GC-MS 53 volatile compounds investigated floating particles that related combustion gases, such saturated alkanes, cycloalkanes, esters, aromatic hydrocarbons. It proves contamination measured matters make them more hazardous health. Viable strategies vehicle traffic-related pollutants mitigation would be reducing occurrence usage modern catalyst filters gas exhausting system.

Language: Английский

---