Pith. sign in

REVIEW 4 major objections 6 minor 44 references

Comparative Analysis of Machine Learning-Based Imputation Techniques for Air Quality Datasets with High Missing Data Rates

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that even with 82.42% of grid-hour PM2.5 values missing, machine-learning classifiers can still identify pollution levels with near-95% accuracy and F1 score.

desk verdict The paper's reported PM2.5 classification accuracy is essentially the majority-class rate, and the features include same-hour PM2.5 readings from nearby stations, so the headline claim is unsupported. read the letter →

arxiv 2412.13966 v2 pith:B4DVJNZA submitted 2024-12-18 cs.LG physics.data-an

Sen Yan , David J. O'Connor , Xiaojun Wang , Noel E. O'Connor , Alan F. Smeaton , Mingming Liu This is my paper · ORCID
classification cs.LGphysics.data-an
keywords airqualityPM2.5classificationdataimputationmissingdiffusionmodelsrandomforestexternalfeaturesspatiotemporal
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that machine-learning imputation can still classify PM2.5 (fine particulate matter) levels accurately when the underlying urban monitoring grid is missing about 82% of its hourly cell readings. Using a merged Dublin dataset of mobile and fixed-station measurements, the authors compare conventional ensemble models, recurrent deep networks, and diffusion models, with and without external traffic, weather, and nearest-station features. Their central result is that a latent diffusion model reaches the highest F1 score of 0.9486 (a balanced measure of precision and recall) while random forest reaches the highest accuracy of 94.82%, both with external features. The claim matters because sparse, multi-source city sensing is common, and it suggests that high missingness alone need not block useful air-quality classification.

What carries the argument

The central object is the processed spatiotemporal dataset: inner Dublin divided into 70 grid cells of 500 m by 500 m, with hourly PM2.5 values averaged from three sensor networks, leaving an 82.42% missing rate. Around this grid the models see temporal features (weekday, hour), one-hot spatial cell codes, and external features (traffic volume, weather variables, and distance and observations from the two nearest fixed stations); PM2.5 is binned into four classes from the European AQI scale and rebalanced with SMOTE. The comparison is carried by two families of classifiers: ensemble models (random forest, XGBoost) and diffusion models (DDPM and LDM), where LDM first encodes input into a latent space, runs the diffusion process there, and decodes back for classification. This shared grid-and-feature setup is what makes the accuracy and F1 comparisons meaningful.

What would settle it

Train the best models on the same data but evaluate on a temporally held-out slice that mimics the systematic gaps, for example masking all night and weekend hours or holding out entire days, and compare accuracy and F1 with the reported random 4:1 split; a substantial drop would falsify the claim that high missingness alone is manageable, and comparing observed versus masked cells using nearest-station data would reveal bias.

Watch

Extended reading notes

Core claim

The central claim is that on a 500 m by 500 m hourly grid of inner-Dublin PM2.5 readings, where 82.42% of grid-hour values are absent after merging three monitoring sources, classification of PM2.5 into four European AQI levels remains highly accurate. With external features (traffic volume, weather, and observations from the two nearest fixed stations), the latent diffusion model achieved the best F1 score of 0.9486 (accuracy 94.26%, precision 95.89%, recall 94.26%), and random forest achieved the best accuracy of 94.82% with F1 0.9440. Diffusion models also produced more balanced per-class results, with notably higher accuracy and AUC for pollution classes above the very-low category. The authors interpret this as evidence that ensemble and diffusion methods, aided by external features, can cope with high missing-data rates in spatiotemporal air-quality datasets.

Load-bearing premise

The load-bearing assumption is that the 82.42% missingness can be treated as random, so patterns learned from observed grid-hour labels generalise to the missing cells; in reality the sensors run on weekday daytime schedules and delivery routes, so absence may correlate with pollution and traffic.

Editorial extensions

If this is right

  • A missing-data rate above 80% does not by itself prevent accurate PM2.5 level classification, provided models are trained on observed labels and given external features.
  • External features substantially improve simpler models: KNN accuracy rose from 81.31% to 88.63% and MLP from 72.06% to 90.38% when traffic, weather, and nearest-station data were added.
  • For applications where missing a high-pollution event is costly, latent diffusion models may be preferable to accuracy-maximising ensembles because they balance per-class recall better.
  • Merging mobile and fixed sensor sources reduces missingness from roughly 90-98% per source to 82.42% and is a useful preprocessing step for sparse urban sensing.
  • The comparative setup can serve as a template for evaluating imputation and classification on other pollutants and cities with sparse monitoring.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper leaves open whether the near-95% numbers transfer to filling the actual missing cells: the evaluation uses randomly held-out observed labels, while the missingness is systematic (vehicle-mounted sensors run mainly on weekdays, during daytime, and along delivery routes). A fairer test would mask entire days or the unobserved night and weekend cells.
  • Because only one raw observation exceeded 110 micrograms per cubic metre and the top AQI classes were merged, the results validate discrimination among low, moderate, and occasionally high pollution, not the full AQI range; alert systems would need rare-event data from longer periods or other cities.
  • The finding suggests a testable extension: feeding diffusion-model class probabilities into route planners for cyclists and pedestrians could turn air-quality classification into exposure-aware navigation, a direction the authors mention as future work.
  • A related open question is whether the external-feature gains reflect genuine causal drivers (traffic and weather) or simply correlations with the sampling schedule; comparing a model trained with shuffled timestamps would separate the two.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. This paper presents a comparative study of machine learning imputation and prediction methods for PM2.5 level classification on a merged Dublin air-quality dataset with an approximately 82% missing rate. The authors grid the data into 500 m cells, merge Google, EPA, and DPD observations, and classify PM2.5 into four AQI-based categories using KNN, RF, XGB, MLP, LSTM, GRU, DDPM, and LDM, with and without external features including traffic, weather, and nearest-station observations. They report that RF achieves the highest accuracy (94.82%) and LDM the highest F1 score (0.9486), and interpret this as evidence that good performance is possible despite high missingness.

Significance. The paper addresses a practically important problem: air quality classification under sparse spatiotemporal coverage. It also provides a broad comparison of model families, including diffusion models, which is of interest to the community. However, the headline results are indistinguishable from a trivial majority-class predictor, and the experimental design does not actually evaluate imputation. As a result, the paper does not currently provide a reliable benchmark or substantiate its claims about handling high missing data rates.

major comments (4)
  1. [Section III.D and Table II] The class counts reported in Section III.D (25,901:1,227:167:17:1) imply that, after merging the last two classes, the majority class constitutes 25,901/27,313 = 94.83% of the data. Table II reports RF accuracy of 94.82% and LDM F1 of 0.9486, which are essentially equal to the accuracy and weighted F1 of a dummy classifier that always predicts the majority class. Because no majority baseline, macro-F1, or per-class metrics are reported for the headline numbers, the paper's central claim that these models achieve near-95% performance is unsupported.
  2. [Table I and Section IV.D] The external feature set used in the 'with external features' (wf) runs includes s_val_1 and s_val_2, which are the same-hour PM2.5 observations from the two nearest stations, i.e., direct measurements of the target variable at nearby locations. Since Table I and Section III.D confirm these are PM2.5 values and the target is the PM2.5 level in a grid cell, the model can largely copy the station readings. Section IV.D attributes large accuracy gains (e.g., KNN from 81.31% to 88.63%) to external features without ablating these station observations from the traffic/weather features, so the claimed benefit of external features is not established.
  3. [Sections III.B and IV.A] The manuscript is framed as an imputation study, but the experiments classify observed labels on a 4:1 random split; missing cells are never imputed or evaluated. In addition, Section III.B documents that Google sensors only operate during daytime on weekdays and DPD sensors follow delivery routes, meaning the 82.42% missingness is systematic rather than random. Training and testing only on observed labels therefore cannot demonstrate performance on missing cells, and the abstract's claim about good performance 'despite a high missing data rate' is not supported by the experiments.
  4. [Section IV.A.1] The train/test split is described only as a 4:1 ratio with no temporal ordering. For hourly time series with strong autocorrelation (Section III.B), a random split can place temporally adjacent observations in both training and test, inflating accuracy. A temporal split or explicit block-wise validation is needed for the reported comparisons to be reliable.
minor comments (6)
  1. [Abstract, Section I, Section III.D] The reported missing rate is 82.42% in the abstract and contributions but 82.32% in Section III.D, and the contribution list gives RF accuracy as 94.70% while Table II reports 94.82%; these inconsistencies should be reconciled.
  2. [Section II.A] 'Deep Learnining' should be 'Deep Learning'.
  3. [Table I] The table header contains typos: 'Relative Hmidity' should be 'Relative Humidity' and 'Average Traffic V olume' should be 'Average Traffic Volume'.
  4. [Section IV.A.1] The phrase 'without two different feature setups' should be 'with two different feature setups'.
  5. [Table II and Section IV.B] Several entries list Recall equal to Accuracy (e.g., RF 0.9482/0.9482), which suggests the metrics are weighted or micro-averaged; the averaging scheme should be stated explicitly, and macro-F1 should be reported.
  6. [Section III.D] The paper states missing values in the label column are filled with -1, but it does not explain whether these entries are excluded from training and testing; this should be clarified.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation chain; the high accuracy / F1 issues are external-validity and leakage concerns, not circularity.

full rationale

This is an empirical benchmarking paper with no analytic derivation or fitted-parameter loop to reduce. The train/test split is explicit (4:1) and SMOTE is applied only to the training set (Section IV.A), so the reported held-out accuracies are not produced by fitting a parameter to the test quantity and then predicting it. The 'wf' vs 'nf' paired results (Table II) provide an external-feature ablation, and the diffusion models are compared against conventional ML and DL baselines, so there is no claim whose premise is the conclusion restated. The strongest concerns — same-hour nearest-station PM2.5 readings (s_val_1, s_val_2) used as input features, the majority class being 25,901/27,313 = 94.83% of the processed data, and the fact that missing cells are never actually imputed or evaluated — are important threats to what the numbers mean, but they are soundness/leakage issues rather than circularity: the paper does not define the prediction in terms of its own output, nor does it rely on a self-citation chain to license its central claim. The only self-citations ([1], [3]) are for background and future-work suggestions and are not load-bearing. Accordingly, under the strict circularity rubric, the derivation is self-contained and the score is 0.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The central result depends on a small number of modeling choices, the most important being the treatment of missingness as ignorable and the inclusion of same-hour nearest-station PM2.5 readings as features. Both are unvalidated assumptions rather than derived results; the free parameters are mostly conventional hyperparameters, none of which is fitted to an external benchmark.

free parameters (5)
  • Class merging rule = High and Very High AQI levels merged into one class (threshold at 55 µg/m3)
    The four-class target is created by combining the top two AQI categories to reduce imbalance (Section III.D). This choice affects the meaning of accuracy and F1.
  • SMOTE sampling configuration = Not specified
    SMOTE is applied to the training set (Section III.D) with no stated oversampling ratio or k-neighbors setting, so the synthetic data distribution is an unreported modeling choice.
  • Train/test split type = Ratio 4:1, split mechanism not stated
    Section IV.A.1 specifies a 4:1 split but not whether it is random or temporal; this determines whether temporal autocorrelation leaks into test metrics.
  • Nearest-station feature selection = Two nearest stations, distance-weighted
    s dis 1, s val 1, s dis 2, s val 2 are engineered from the two closest fixed stations (Section III.D, Table I). The choice of two stations and the distance weighting is ad hoc and strongly influences performance.
  • KNN k and RF n_estimators = k=4, n_estimators=10
    These hyperparameters are stated without tuning or sensitivity analysis (Section IV.A.1), yet they directly determine the reported results.
assumptions (3)
  • domain assumption Missingness is ignorable
    The paper trains on non-missing labels and evaluates on held-out non-missing labels without modeling the missingness mechanism or comparing observed to missing cells (Section III.A, III.D). The data description shows systematic absence, making this assumption load-bearing.
  • domain assumption Hourly averaging within 500m grid cells yields representative PM2.5 levels
    Google, EPA, and DPD measurements are merged by hourly mean per cell (Section III.D), implicitly assuming the sparse mobile samples represent the whole cell.
  • domain assumption AQI thresholds from an external API define meaningful classes
    The label categories come from the OpenWeather European AQI scale (Section III.D, [41]) and are not validated against local health outcomes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Comparative Analysis of Machine Learning-Based Imputation Techniques for Air Quality Datasets with High Missing Data Rates." pith.science (2026). https://pith.science/paper/B4DVJNZA

@misc{pith2026241213966,
  author       = {Pith},
  title        = {Pith review of: Comparative Analysis of Machine Learning-Based Imputation Techniques for Air Quality Datasets with High Missing Data Rates},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B4DVJNZA}},
  note         = {Machine review of arXiv:2412.13966}
}
read the original abstract

Urban pollution poses serious health risks, particularly in relation to traffic-related air pollution, which remains a major concern in many cities. Vehicle emissions contribute to respiratory and cardiovascular issues, especially for vulnerable and exposed road users like pedestrians and cyclists. Therefore, accurate air quality monitoring with high spatial resolution is vital for good urban environmental management. This study aims to provide insights for processing spatiotemporal datasets with high missing data rates. In this study, the challenge of high missing data rates is a result of the limited data available and the fine granularity required for precise classification of PM2.5 levels. The data used for analysis and imputation were collected from both mobile sensors and fixed stations by Dynamic Parcel Distribution, the Environmental Protection Agency, and Google in Dublin, Ireland, where the missing data rate was approximately 82.42%, making accurate Particulate Matter 2.5 level predictions particularly difficult. Various imputation and prediction approaches were evaluated and compared, including ensemble methods, deep learning models, and diffusion models. External features such as traffic flow, weather conditions, and data from the nearest stations were incorporated to enhance model performance. The results indicate that diffusion methods with external features achieved the highest F1 score, reaching 0.9486 (Accuracy: 94.26%, Precision: 94.42%, Recall: 94.82%), with ensemble models achieving the highest accuracy of 94.82%, illustrating that good performance can be obtained despite a high missing data rate.

Figures

Figures reproduced from arXiv: 2412.13966 by the authors.

Figure 1
Figure 1. Comparison of hourly average PM2.5 values in 3 datasets, with time series decomposition of DPD data as an example [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Spatial distribution of PM2.5 data in three datasets. area with significant overlap, indicating more extensive data collection across the city. The different point sizes in the DPD and Google datasets indicate fluctuations in pollution intensity throughout our focus area and provide a more comprehensive view of air quality dynamics. D. Data Processing In this study, we divided the inner Dublin area into a grid of 50… view at source ↗
Figure 3
Figure 3. Confusion matrix and ROC curves for RF and LDM. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 44 canonical work pages

  1. [1]

    Breathing green: Maximising health and environmental benefits for active transportation users leveraging large scale air quality data,

    S. Yan, S. Zhu, J. B. Fernandez, E. A. S ´anchez, Y . Gu, N. E. O’Connor, D. O’Connor, and M. Liu, “Breathing green: Maximising health and environmental benefits for active transportation users leveraging large scale air quality data,” in 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC) . IEEE, Sep. 2023

    work page 2023

  2. [2]

    Particulate matter air pollution is a significant risk factor for cardiovascular disease,

    R. J. Henning, “Particulate matter air pollution is a significant risk factor for cardiovascular disease,” Current Problems in Cardiology , vol. 49, no. 1, p. 102094, Jan. 2024

    work page 2024

  3. [3]

    Parking behaviour analysis of shared e-bike users based on a real-world dataset - A case study in Dublin, Ireland,

    S. Yan, M. Liu, and N. E. O’Connor, “Parking behaviour analysis of shared e-bike users based on a real-world dataset - A case study in Dublin, Ireland,” in 2022 IEEE 95th Vehicular Technology Conference: (VTC2022-Spring). IEEE, Jun. 2022

    work page 2022

  4. [4]

    PM2.5 Exposure and asthma development: The key role of oxidative stress,

    K. Liu, S. Hua, and L. Song, “PM2.5 Exposure and asthma development: The key role of oxidative stress,” Oxidative Medicine and Cellular Longevity, vol. 2022, p. 1–12, Apr. 2022

    work page 2022

  5. [5]

    Relation between pm2.5 pollution and covid-19 mor- tality in western europe for the 2020–2022 period,

    J.-B. Renard, J. Surcin, I. Annesi-Maesano, G. Delaunay, E. Poincelet, and G. Dixsaut, “Relation between pm2.5 pollution and covid-19 mor- tality in western europe for the 2020–2022 period,” Science of The Total Environment, vol. 848, p. 157579, Nov. 2022

    work page 2020

  6. [6]

    Temporal evolution of pm2.5 levels and covid-19 mortality in europe for the 2020–2022 period,

    J.-B. Renard, J. Surcin, I. Annesi-Maesano, and E. Poincelet, “Temporal evolution of pm2.5 levels and covid-19 mortality in europe for the 2020–2022 period,” Atmosphere, vol. 14, no. 8, p. 1222, Jul. 2023

    work page 2020

  7. [7]

    A comparative analysis for air quality estimation from traffic and meteorological data,

    E. Arnaudo, A. Farasin, and C. Rossi, “A comparative analysis for air quality estimation from traffic and meteorological data,” Applied Sciences, vol. 10, no. 13, p. 4587, Jul. 2020

    work page 2020

  8. [8]

    Machine learning for air quality prediction using meteorological and traffic related features,

    I. Gryech, M. Ghogho, H. Elhammouti, N. Sbihi, and A. Kobbane, “Machine learning for air quality prediction using meteorological and traffic related features,” Journal of Ambient Intelligence and Smart Environments, vol. 12, no. 5, p. 379–391, Sep. 2020

    work page 2020

Show all 44 references
  1. [9]

    Time series analysis and forecasting of air quality index of Dhaka City of Bangladesh,

    S. R. Sakib, K. N. Sara, M. T. Hossain Rasel, M. M. Islam Asif, M. A. Hasan Nahid, M. S. Rahman, M. F. Mridha, and A. Islam, “Time series analysis and forecasting of air quality index of Dhaka City of Bangladesh,” in 2023 IEEE World AI IoT Congress (AIIoT) . IEEE, Jun. 2023

    2023

  2. [10]

    Air-pollution prediction in smart city, deep learning approach,

    A. Bekkar, B. Hssina, S. Douzi, and K. Douzi, “Air-pollution prediction in smart city, deep learning approach,” Journal of Big Data, vol. 8, no. 1, Dec. 2021

    2021

  3. [11]

    A deep learning approach for prediction of air quality index in a metropolitan city,

    R. Janarthanan, P. Partheeban, K. Somasundaram, and P. Navin Elam- parithi, “A deep learning approach for prediction of air quality index in a metropolitan city,” Sustainable Cities and Society , vol. 67, p. 102720, Apr. 2021

    2021

  4. [12]

    Modeling air quality prediction using a deep learning approach: Method optimization and evaluation,

    W. Mao, W. Wang, L. Jiao, S. Zhao, and A. Liu, “Modeling air quality prediction using a deep learning approach: Method optimization and evaluation,” Sustainable Cities and Society , vol. 65, p. 102567, Feb. 2021

    2021

  5. [13]

    Quantifying uncertainty: Air quality forecasting based on dynamic spatial-temporal denoising diffusion probabilistic model,

    K. Chen, G. Li, H. Li, Y . Wang, W. Wang, Q. Liu, and H. Wang, “Quantifying uncertainty: Air quality forecasting based on dynamic spatial-temporal denoising diffusion probabilistic model,” Environmental Research, vol. 249, p. 118438, May 2024

    2024

  6. [14]

    Short-term wind power scenario generation based on conditional latent diffusion models,

    X. Dong, Z. Mao, Y . Sun, and X. Xu, “Short-term wind power scenario generation based on conditional latent diffusion models,” IEEE Transactions on Sustainable Energy , vol. 15, no. 2, p. 1074–1085, Apr. 2024

    2024

  7. [15]

    Forecasting the exceedances of PM2.5 in an urban area,

    S.-A. Logothetis, G. Kosmopoulos, O. Panagopoulos, V . Salamalikis, and A. Kazantzidis, “Forecasting the exceedances of PM2.5 in an urban area,” Atmosphere, vol. 15, no. 5, p. 594, May 2024

    2024

  8. [16]

    A Machine Learning Approach to Predict Air Quality in California,

    M. Castelli, F. M. Clemente, A. Popovi ˇc, S. Silva, and L. Vanneschi, “A Machine Learning Approach to Predict Air Quality in California,” Complexity, vol. 2020, p. 1–23, Aug. 2020

    2020

  9. [17]

    A hybrid deep learning technology for PM2.5 air quality forecasting,

    Z. Zhang, Y . Zeng, and K. Yan, “A hybrid deep learning technology for PM2.5 air quality forecasting,” Environmental Science and Pollution Research, vol. 28, no. 29, p. 39409–39422, Mar. 2021

    2021

  10. [18]

    Spatio-attention embedded recurrent neural network for air quality prediction,

    Y . Huang, J. J.-C. Ying, and V . S. Tseng, “Spatio-attention embedded recurrent neural network for air quality prediction,” Knowledge-Based Systems, vol. 233, p. 107416, Dec. 2021

    2021

  11. [19]

    Multi-scale spatiotemporal graph convolution network for air quality prediction,

    L. Ge, K. Wu, Y . Zeng, F. Chang, Y . Wang, and S. Li, “Multi-scale spatiotemporal graph convolution network for air quality prediction,” Applied Intelligence, vol. 51, no. 6, p. 3491–3505, Nov. 2020

    2020

  12. [20]

    PM10 and PM2.5 real-time prediction models using an interpolated convolutional neural network,

    S. Chae, J. Shin, S. Kwon, S. Lee, S. Kang, and D. Lee, “PM10 and PM2.5 real-time prediction models using an interpolated convolutional neural network,” Scientific Reports, vol. 11, no. 1, Jun. 2021

    2021

  13. [21]

    Regional air quality forecasting using spatiotemporal deep learning,

    S. Abirami and P. Chitra, “Regional air quality forecasting using spatiotemporal deep learning,” Journal of Cleaner Production, vol. 283, p. 125341, Feb. 2021

    2021

  14. [22]

    A novel multi-pollutant space-time learning network for air pollution inference,

    J. Song and M. E. Stettler, “A novel multi-pollutant space-time learning network for air pollution inference,” Science of The Total Environment , vol. 811, p. 152254, Mar. 2022

    2022

  15. [23]

    Spatiotemporal deep learning model for citywide air pollution interpolation and prediction,

    V .-D. Le, T.-C. Bui, and S.-K. Cha, “Spatiotemporal deep learning model for citywide air pollution interpolation and prediction,” in 2020 IEEE International Conference on Big Data and Smart Computing (BigComp). IEEE, Feb. 2020

    2020

  16. [24]

    Linking of open and private data in dataspace: A case study of air quality monitoring and forecasting,

    A. Acquier, A. Donald, E. Curry, I. Ullah, and U. ul Hassan, “Linking of open and private data in dataspace: A case study of air quality monitoring and forecasting,” in Proceedings of the Second International Workshop on Semantics in Dataspaces (SDS 2024) co-located with the 2...

    2024

  17. [25]

    Quantifying road traffic impact on air quality in urban areas: A covid19- induced lockdown analysis in Italy,

    G. Gualtieri, L. Brilli, F. Carotenuto, C. Vagnoli, A. Zaldei, and B. Gioli, “Quantifying road traffic impact on air quality in urban areas: A covid19- induced lockdown analysis in Italy,” Environmental Pollution, vol. 267, p. 115682, Dec. 2020

    2020

  18. [26]

    Optimized feature selection for air quality index forecasting using GPR and SARIMA models,

    S. Veera Manikandan, Y . Abilash, S. Hari Prasanth, J. Alfred Daniel, and R. Santhosh, “Optimized feature selection for air quality index forecasting using GPR and SARIMA models,” in 2023 International Conference on Inventive Computation Technologies (ICICT) . IEEE, Apr. 2023

    2023

  19. [27]

    Comparison of missing data imputation methods in time series forecasting,

    H. Ahn, K. Sun, and K. Pio Kim, “Comparison of missing data imputation methods in time series forecasting,” Computers, Materials & Continua, vol. 70, no. 1, p. 767–779, 2022

    2022

  20. [28]

    A graph-based approach for missing sensor data imputation,

    X. Jiang, Z. Tian, and K. Li, “A graph-based approach for missing sensor data imputation,” IEEE Sensors Journal , vol. 21, no. 20, p. 23133–23144, Oct. 2021

    2021

  21. [29]

    Random forest spatial interpolation,

    A. Sekuli ´c, M. Kilibarda, G. B. Heuvelink, M. Nikoli ´c, and B. Bajat, “Random forest spatial interpolation,” Remote Sensing, vol. 12, no. 10, p. 1687, May 2020

    2020

  22. [30]

    Spatiotempo- ral high-resolution imputation modeling of aerosol optical depth for investigating its full-coverage variation in China from 2003 to 2020,

    Q. He, W. Wang, Y . Song, M. Zhang, and B. Huang, “Spatiotempo- ral high-resolution imputation modeling of aerosol optical depth for investigating its full-coverage variation in China from 2003 to 2020,” Atmospheric Research, vol. 281, p. 106481, Jan. 2023

    2003

  23. [31]

    Machine learning based approaches for imputation in time series data and their impact on forecasting,

    M. Saad, M. Chaudhary, F. Karray, and V . Gaudet, “Machine learning based approaches for imputation in time series data and their impact on forecasting,” in 2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC) . IEEE, Oct. 2020

    2020

  24. [32]

    An air quality forecasting model based on improved convnet and RNN,

    B. Wang, W. Kong, and P. Zhao, “An air quality forecasting model based on improved convnet and RNN,” Soft Computing, vol. 25, no. 14, p. 9209–9218, May 2021

    2021

  25. [33]

    Air pollution prediction using lstm deep learning and metaheuristics algorithms,

    G. I. Drewil and R. J. Al-Bahadili, “Air pollution prediction using lstm deep learning and metaheuristics algorithms,” Measurement: Sensors , vol. 24, p. 100546, Dec. 2022

    2022

  26. [34]

    Google airview data - Dublin City,

    Google, “Google airview data - Dublin City,” 2023, accessed: 2024- 09-15. [Online]. Available: https://data.smartdublin.ie/dataset/google-a irview-data-dublin-city

    2023

  27. [35]

    EPA open data,

    Environmental Protection Agency (EPA) Ireland, “EPA open data,” 2024, accessed: 2024-09-15. [Online]. Available: https://data.epa.ie/ap i-list/epr-open-data

    2024

  28. [36]

    Spatial distribution of pm2.5 mass and number concentrations in paris (france) from the pollutrack network of mobile sensors during 2018–2022,

    J.-B. Renard, E. Poincelet, I. Annesi-Maesano, and J. Surcin, “Spatial distribution of pm2.5 mass and number concentrations in paris (france) from the pollutrack network of mobile sensors during 2018–2022,” Sensors, vol. 23, no. 20, p. 8560, Oct. 2023

    2018

  29. [37]

    High resolution mapping of pm2.5 concentrations in paris (france) using mobile pollutrack sensors network in 2020,

    J.-B. Renard and C. Marchand, “High resolution mapping of pm2.5 concentrations in paris (france) using mobile pollutrack sensors network in 2020,” Atmosphere, vol. 12, no. 5, p. 529, Apr. 2021

    2020

  30. [38]

    DCC SCATS detector volume (jan-jun 2022),

    Dublin City Council, “DCC SCATS detector volume (jan-jun 2022),” 2022, accessed: 2024-09-15. [Online]. Available: https: //data.smartdublin.ie/dataset/dcc-scats-detector-volume-jan-jun-2022

    2022

  31. [39]

    DCC SCATS detector volume (jul-dec 2022),

    Dublin City Council, “DCC SCATS detector volume (jul-dec 2022),” 2022, accessed: 2024-09-15. [Online]. Available: https: //data.smartdublin.ie/dataset/dcc-scats-detector-volume-jul-dec-2022

    2022

  32. [40]

    Historical climate data,

    Met ´Eireann, “Historical climate data,” 2024, accessed: 2024-09-15. [Online]. Available: https://www.met.ie/climate/available-data/historical -data

    2024

  33. [41]

    Air pollution index levels - Europe,

    OpenWeather, “Air pollution index levels - Europe,” 2024, accessed: 2024-09-15. [Online]. Available: https://openweathermap.org/air-polluti on-index-levels#europe

    2024

  34. [42]

    Geneva: World Health Organization,

    World Health Organization, WHO Global Air Quality Guidelines: Particulate Matter (PM2.5 and PM10), Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide . Geneva: World Health Organization,

  35. [43]

    SMOTE: Synthetic minority over-sampling technique,

    N. V . Chawla, K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer, “SMOTE: Synthetic minority over-sampling technique,” Journal of Artificial Intelligence Research , vol. 16, pp. 321–357, Jun. 2002

    2002

  36. [2021]

    Available: https://www.who.int/publications/i/item/9789 240034228

    [Online]. Available: https://www.who.int/publications/i/item/9789 240034228

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.