Figure - available from: PLOS One
This content is subject to copyright.
Source publication
Today, China is facing a very serious issue of Air Pollution due to its dreadful impact on the human health as well as the environment. The urban cities in China are the most affected due to their rapid industrial and economic growth. Therefore, it is of extreme importance to come up with new, better and more reliable forecasting models to accurate...
Similar publications
Purpose of this paper: Nowadays, the received signal strength based WiFi indoor positioning, a low-cost indoor positioning approach, has attracted significant attention from both academia and industry. The weighted squared Euclidean distance between points in signal space is introduced in this paper to improve the performance of the WiFi indoor pos...
![Fig. 4. Predicted versus actual energy demand in May 2011 [11]](profile/Roger-Anderson-4/publication/325651780/figure/fig3/AS:635263910948865@1528470127836/Predicted-versus-actual-energy-demand-in-May-2011-11_Q320.jpg)
A high percentage of newly-constructed commercial office buildings experience energy consumption that exceeds specifications and system failures after being put into use. This problem is even worse for older buildings. We present a new approach, 'predictive building energy optimization', which uses machine learning (ML) and automated online evaluat...
Landslide displacement prediction is an essential component for developing landslide early warning systems. In the Three Gorges Reservoir area (TGRA), landslides experience step-like deformations (i.e., periods of stability interrupted by abrupt accelerations) generally from April to September due to the influence of precipitation and reservoir sch...
Resource Management in the network function virtualization (NFV) environment is a challenging task. The continuously varying demands of virtual network functions (VNF) call for dynamic algorithms to efficiently scale the allocated resources and meet fluctuating needs. In this context, studying the behavior of a VNF as a function of its environment...
Here we present a method for prediction of HPLC retention times based on support vector regression. In contrast to existing prediction methods for DNA, our method takes the secondary structure of DNA into account. The method is also well suited for retention time prediction of peptides.
Citations
... Collaborative and multi-dimensional models Liu et al. (2017) introduced a multi-dimensional collective support vector regression (SVR) model for AQI prediction in the Beijing-Tianjin-Shijiazhuang region. Their research emphasized the significance of geographical and meteorological factors in improving prediction accuracy, demonstrating that multi-city data collaboration can effectively reduce prediction errors (Liu et al., 2017). ...
... Collaborative and multi-dimensional models Liu et al. (2017) introduced a multi-dimensional collective support vector regression (SVR) model for AQI prediction in the Beijing-Tianjin-Shijiazhuang region. Their research emphasized the significance of geographical and meteorological factors in improving prediction accuracy, demonstrating that multi-city data collaboration can effectively reduce prediction errors (Liu et al., 2017). This model exemplifies the benefits of considering spatial correlations in air quality forecasting. ...
India, the most populous country in the world, ranks as the seventh largest by area. As per IQAir reports, in 2024, India was the fifth most polluted country, preceded by Chad, Congo, Bangladesh, and Pakistan, based on Air Quality Index (AQI) values. This study aims to predict air quality in Pune, Maharashtra, using an AI-driven data-centric approach. The dataset, obtained from sources such as Kaggle, CPCB, and WHO, comprises 3,170 records covering fifteen key factors influencing AQI, including SO₂, NOx, RSPM, precipitation, maximum and minimum temperature, sun hours, UV index, wind gust, humidity, pressure, average temperature, and wind speed. Data spanning nineteen years (2006–2024) is utilized to develop the predictive model, with records from 2006–2019 used for training and testing, while data from 2020–2024 is reserved for validation. This research proposes Linear Regression (LR) as a machine learning approach, achieving an R-value of 0.9611. The LR model's performance metrics include an RMSE of 21.4079, MAPE of 7.8945%, and MAE of 13.5884. The developed model can assist in forecasting air quality for urban residents, contributing to public health protection. Furthermore, it can aid in identifying effective mitigation strategies and operational measures to enhance air quality.
... Green's Combined Index (GCI) [18], Ontario Air Pollution Index (O-API) [19], PINDEX [20], Oak Ridge Air Quality Index (OR-AQI) [21], New Jersey Air Quality Index (NJ-AQI) [22], Observer-Based Air Quality Index (OB-AQI) [23], and Most Undesirable Respirable Contaminants Index (MURC) [24] are some of the earlier methods to evaluate the air quality in built-up areas. Subsequently, some of the most recent developments include the Fuzzy-based Air Quality Index (F-AQI) [25], coupled Artificial Neural Network [26], Support Vector Regression [27]. In India, a quantification effort on behalf of the National Air Quality Monitoring Program was initiated much later, just after 1984. ...
... The COVID-19 pandemic has led to global attention on air pollution, with significant reductions in pollutants during lockdowns worldwide. For example, studies have shown a 30% reduction in NO 2 , NH 3 , and a 25% decrease in carbon emissions in China [27,45]. Research by Watts and Kommenda [46] and Cadotte [47] also reported temporary cuts in air pollutants due to industrial shutdowns in major cities. ...
Utilizing data from real-time monitoring stations of the Central Pollution Control Board (CPCB), which encompass pollution activities in the region, including stubble burning and other anthropogenic sources, this study evaluates the National Air Quality Index (NAQI). The analysis examines spatial variations in seven key air pollutants—particulate matter (PM2.5 and PM10) and gaseous pollutants (NH3, NO2, O3, CO, and SO2)—across Punjab, India during three distinct phases: pre-lockdown, complete lockdown, and post-lockdown, providing insights into the shifts in air quality influenced by both natural and human-induced factors. The results showed that air quality was significantly improved during the lockdown. PM2.5, PM10, and NH3 saw the most substantial reductions (40–60%). The absolute drop in PM2.5, PM10, and NH3 was noted during peak morning traffic hours (08–10 Hrs) and late evening traffic hours (20–24 Hrs); however, the percentage reduction remained nearly consistent throughout the day. SO2 and O3 had a mixed pattern of increase and decline, with a modest increase at some sites and a considerable decrease at others. In the central region, there was a significant drop in NO2 and CO (50–60%). The decrease in PM2.5, PM10, NH3, NO2, and increase in O3 was related to population density. As a result, Punjab’s air quality has improved, with a significant drop in major pollutants; nonetheless, some regions have seen an increase in SO2 and O3. Changes in emissions, meteorology and atmospheric chemistry documented during the lockdown necessitate further examination.
... A trained neural network model can predict air quality for a future period of time based on factors such as geographic location, environment, time of day, and meteorological parameters [21][22][23]. In addition, machine learning algorithms such as Random Forest [24][25][26], Gaussian Process Regression (GPR) [27] and Support Vector Regression (SVR) [28,29] are widely used for air quality prediction. These algorithms analyze historical data and build models to predict future air quality conditions. ...
As a key tool for real-time monitoring of air pollutant concentrations, the chemical sensor, the core component of the low-cost Air Quality Monitor (AQM), is susceptible to a variety of factors during the measurement process, leading to errors in the measurement data. To enhance the measurement accuracy of chemical sensors, this paper presents a calibration method based on the PCR-GPR model. This method not only effectively enhances the measurement accuracy of chemical sensors, but also combines the interpretability of traditional statistical models with the high-precision characteristics of Gaussian Process Regression (GPR) models. First, we perform Principal Component Analysis (PCA) on the measurement data of the AQM to solve the multicollinearity problem. Through PCA, we successfully extracted 8 principal components, which not only contained 95% of the information in the original data, but also effectively eliminated the correlation between the variables, providing a more robust data base for subsequent modeling. Subsequently, we established a Principal Component Regression (PCR) model using the concentration of pollutants measured by the national monitoring station as the dependent variable and the 8 principal components extracted above as the independent variables. The PCR model can effectively extract the linear relationship between the independent and dependent variables, providing a linear part of the explanation for the calibration process. However, there are often complex nonlinear relationships between pollutant concentrations and AQM measurements. To capture these nonlinear relationships, we further established a GPR model with the residuals of the PCR model as the dependent variable and the measurement data of the AQM as the independent variable. By combining the PCR model and the GPR model, we obtained the final PCR-GPR calibration model. It is worth mentioning that this study adopted the time series cross-validation method for data grouping, an innovative approach that is more aligned with real-world scenarios and adequately captures the seasonal variations in pollutant concentrations. The experimental results show that the model exhibits excellent performance on several evaluation metrics and can calibrate the chemical sensor well, improving its measurement accuracy by 16.94% ~ 82.01%.
... recasting air quality, particularly during periods of heavy pollution Ma et al., 2020;Pan, 2018). The SVR model performs well in identifying key samples and filtering extreme samples, which enhances the generalization capability and shows strong performance in predicting lower PM 2.5 concentrations in Haikou (Ke, Gong, He, Zhang, Cui, et al., 2022;B.-C. Liu et al., 2017;Sánchez et al., 2011). In addition, the performance of the three stacking models is also very good compared to the original CUACE model, but their evaluation results are still inferior to the XGBoost model and SVR model, which can be related to the combination of base-models chosen. ...
An automated air quality forecasting system (AI‐Air) was developed to optimize and improve air quality forecasting for different typical cities, combined with the China Meteorological Administration Unified Atmospheric Chemistry Environmental Model (CUACE), and used in a typical inland city of Zhengzhou and a coastal city of Haikou in China. The performance evaluation results show that for the PM2.5 forecasts, the correlation coefficient (R) is increased by 0.07–0.13, and the mean error (ME) and root mean square error (RMSE) is decreased by 3.2–3.5 and 3.8–4.7 μg/m³. Similarly, for the O3 forecasts, the R value is improved by 0.09–0.44, and ME and RMSE values are reduced by 7.1–22.8 and 9.0–25.9 μg/m³, respectively. Case analyses of operational forecasting also indicate that the AI‐Air system can significantly improve the forecasting performance of pollutant concentrations and effectively correct underestimation, or overestimation phenomena compared to the CUACE model. Additionally, explanatory analyses were performed to assess the key meteorological factors affecting air quality in cities with different topographic and climatic conditions. The AI‐Air system highlights the potential of AI techniques to improve forecast accuracy and efficiency, and with promising applications in the field of air quality forecasting.
... To reduce the prediction errors of ML algorithms, prediction results can be improved based on the Support Vector Regression (SVR) model utilizing inputs of multidimensional AQI and weather conditions of multiple cities [17]. ...
Air pollution is a critical environmental problem driven by urbanization and industrialization. Time-series forecasting using previous methods is difficult because models must account for seasonal changes, day-to-day changes, and emergencies that can rapidly affect air quality. Therefore, existing approaches struggle to predict these fluctuations. This research addresses this issue by proposing a material-focused method of air quality prediction using machine learning techniques. The proposed model incorporates feature selection using MS-ANFIS-FS and classification using Unet-RNN (Unet Optimized Recurrent Neural Network). The model focuses on analyzing pollutant interactions with material surfaces, improving prediction accuracy by considering the role of materials in pollutant dispersion and absorption. The Successive Feature Defect Scaling Rate (SFDSR) and Auto-Regressive Integrated Moving Average (ARIMA) methods detect variance dependencies in air quality data. These methods enable the model to identify material traits influencing pollution levels, yielding more accurate results for pollutants like PM2.5 and NO2. The findings demonstrate the critical importance of material properties in environmental management and show how material-based interventions can effectively reduce air pollution. This model has the potential to facilitate real-time pollution monitoring and support the development of sustainable air quality management strategies.
... The machine learning method is worthy of recognition in air quality prediction, as it overcomes the non-linear barriers and uncertainties of air pollution. Recent studies [80][81][82] have shown that the model's accuracy in predicting other key components of air, such as CO and NO x , remains limited. Additionally, these models are less accurate in predicting extreme concentrations of atmospheric pollutants. ...
Faced with increasingly serious environmental problems, scientists have conducted extensive research, among which the importance of air quality prediction is becoming increasingly prominent. This article briefly reviews the utilization of geographic artificial intelligence (AI) in air pollution. Firstly, this paper conducts a literature metrology analysis on the research of geographical AI used in air pollution. That is, 607 documents are retrieved from the Web of Science (WOS) using appropriate keywords, and literature metrology analysis is conducted using Citespace to summarize research hotspots and frontier countries in this field. Among them, China plays a constructive role in the fields of geographic AI and air quality research. The data characteristics of Earth science and the direction of AI utilization in the field of Earth science were proposed. It then quickly expanded to investigate and research air pollution. In addition, based on summarizing the current status of Artificial Neural Network (ANN), Recurrent Neural Network (RNN), and hybrid neural network models in predicting air quality (mainly PM2.5), this article also proposes areas for improvement. Finally, this article proposes prospects for future research in this field. This study aims to summarize the development trends and research hotspots of the utilization of geographic AI in the prediction of air quality, as well as prediction methods, to provide direction for future research.
... Classical data-driven approaches like statistical methods (Briggs, 2007;Wang et al., 2009;Rekhi et al., 2019) and machine learning algorithms often use predetermined functions for input-output relationships, which may fail in complex systems and long-term air quality predictions due to their inability to capture implicit relationships. Machine learning techniques such as SVR (Liu et al., 2017), ANNs (Mao et al., 2017), and random forest (Yu et al., 2016) incorporate nonlinearities to better represent complex systems but overlook regional data correlations and also lacks a focus on domain knowledge. ...
... They observed that among the five machine learning algorithms, the stacking ensemble and AdaBoost outperformed other approaches. Liu et al. [20] predicted AQI in the three cities of the Jingjinji Region, China based on the SVR model. They found that SVR increased prediction accuracy and was less prone to overfitting compared to ANN. ...
Currently, Bangladesh is among the countries with the highest levels of air pollution in the world. Major cities of Bangladesh experience severe air pollution, significantly impacting public health and the environment. The Air Quality Index (AQI) serves as a convenient measure to assess air quality in specific locations. Accurate AQI forecasting is crucial as it allows policymakers to prepare and implement strategies to mitigate air pollution, thereby safeguarding human health and the ecosystem. The primary aim of this paper is to develop an effective model capable of forecasting AQI on a medium-term basis for four selected locations in Bangladesh. To achieve this, we propose a Bidirectional Gated Recurrent Unit (BiGRU)-based model to forecast the AQI for each location up to seven days in advance (i.e., medium-term AQI prediction). We collected daily AQI data from four distinct locations in Bangladesh from March 1, 2014, to December 31, 2020. To validate the performance of the model, we used various statistical measures, including Root Mean Square Error (RMSE), Normalized Root Mean Square Error (nRMSE), Mean Absolute Percentage Error (MAPE), and the Coefficient of Determination (R2) score. We also compared the performance of the proposed model with three related models: Gated Recurrent Unit (GRU), Long Short-term Memory (LSTM), and Convolutional Neural Network (CNN). Simulation results demonstrate that the BiGRU model outperforms other approaches, showcasing an RMSE ranging from 27.42 to 43.27 for medium-term AQI prediction.
... SARIMA is selected as a classical model capable of handling time series data with complex seasonal and trend patterns [18]. SVR is preferred for its capacity to handle nonlinear relationships in data, often encountered in air pollution prediction contexts [19]. LSTM is chosen for its capability to capture complex temporal patterns and nonlinear relationships within time series data, making it suitable for modeling air pollution influenced by multiple factors [20]. ...
... SARIMA performs well with complex seasonal and trend patterns but struggles with data exhibiting unstable or changing trends [18]. SVR is adept at handling nonlinear relationships but can be sensitive to parameter tuning and computationally intensive [19]. LSTM, while effective in capturing complex temporal patterns, is prone to overfitting and requires substantial data for effective training [20]. ...
In recent years, data science analysis, particularly time series predictions, has been widely employed across various industrial sectors. However, time series data presents high complexity, especially in seasonal patterns such as monthly, daily, or hourly fluctuations. Irregular fluctuations and external factors increasingly challenge accurate predictions. Therefore, this research proposes a hybrid approach combining SVR-SARIMA, SVR-Prophet, LSTM-SARIMA, and LSTM-Prophet to enhance time series prediction accuracy. This study followed the OSEMN methodology approach: gathering data, cleaning data, exploring data, developing models, and interpreting crucial aspects of problem-solving. Seasonal effect predictions indicated a rise in SO2 and NO2 during dry and rainy seasons until the next two years (average daily increments of 0.0831 μg/m3 for SO2 and 0.0516 μg/m3 for NO2). Estimates suggest a decrease in the order of three particles. The evaluation showed that the SVR model performed better compared to the other three models (RMSE 7.765, MAE 5.477, and MAPE 0.261). The best-performing hybrid model was LSTM-Prophet (99.74% accuracy) with RMSE 12.319, MAE 12.057, and MAPE 0.259 values.
... Based on the multidimensional air quality information and meteorological conditions of Beijing, Tianjin and Shijiazhuang, support vector regression was used by Liu et al. to develop a new collaborative prediction model for predicting the air quality index of Chinese cities. The results show that the Mean Absolute Percentage Error (MAPE) of the multi-city multidimensional regression decreases when there is a strong interaction and correlation between the air quality characteristic attributes and the air quality index (Liu et al., 2017). ...
... The SRA-SVR combined model has a good calibration for PM 2.5 concentrations measured by the miniature sensors, and it needs to be evaluated whether it can also have a good calibration for other pollutants as well. MAE, MAPE and RMSE are used in this study to quantitatively evaluate the performance of each model (Liu et al., 2017;Ratkovic et al., 2023). As can be seen in Tables 4-6, except for the MAPE of the SRA model for SO 2 , the miniature air quality monitor has the maximum of the rest of the indicators, indicating that its measurements need to be calibrated. ...
Effective calibration of miniature air quality monitor measurements is an important task to ensure accurate measurements and guarantee sustainable air quality. The aim of this study is to calibrate the measurement data of miniature air quality monitors using Stepwise Regression Analysis and Support Vector Regression (SRA-SVR) combined model. Firstly, a stepwise regression analysis model is used to find a linear relationship between the measured data from the miniature air quality monitor and the air pollutant concentration. Secondly, support vector regression is used to extract the non-linear relationships which affect the pollutant concentrations hidden in the residuals of the stepwise regression analysis model. Finally, the residual calibration values of the SVR model outputs are added to the SRA model outputs to obtain the final outputs of the SRA-SVR combined model for the pollutants. Mean absolute error, relative mean absolute percent error and root mean square error are used to compare the effectiveness of the SRA-SVR combined model and some other commonly used statistical models for the calibration of miniature air quality monitors. The results show that the SRA-SVR combination model performs optimally on both the training and test sets, regardless of which pollutant and which indicator. The SRA-SVR combined model not only has the advantages of the SRA model’s strong interpretability and the SVR model’s high accuracy, but also has higher accuracy than the single model. By using this model to calibrate the measurements of the miniature air quality monitor, its accuracy can be improved by 61.33%–87.43%.
