Pith. sign in

REVIEW 4 major objections 5 minor 38 references

Spatio-Temporal Forecasting of PM2.5 via Spatial-Diffusion guided Encoder-Decoder Architecture

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read AGNN_GRU: a wind-aware graph plus attention in a GRU encoder-decoder forecasts PM2.5 more accurately than GRU-only or simpler GNN variants on dense station networks.

desk verdict Useful new Bihar PM2.5 dataset and a clean ablation, but the evaluation only compares against self-ablations, so the claimed superiority over published models is not established. read the letter →

arxiv 2412.13935 v1 pith:3VIQ6P7C submitted 2024-12-18 cs.LG cs.AI

classification cs.LGcs.AI
keywords PM2.5forecastingspatio-temporalgraphneuralnetworksencoder-decodergatedrecurrentunitsLuongattentionspatialdiffusionairqualitymonitoring
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 proposes AGNN_GRU, a model for forecasting PM2.5 concentrations simultaneously at many monitoring locations. It is a sequence-to-sequence gated recurrent unit (GRU) model whose encoder adds a graph neural network built on a wind-aware directed graph, and whose decoder uses an attention mechanism that rereads the encoder's history. The central claim is that this combination captures spatial diffusion across locations and short- and long-range temporal dependence better than a plain GRU, a GCN or GraphConv variant, or the same model without attention. That claim is tested on a year of data from 511 low-cost sensors covering Bihar, India, and on four years of data from 181 polluted Chinese cities, with lower error and higher rank correlation reported for most settings. If true, the contribution is a practical forecasting method that needs no meteorological forecasts for the future window, only historical observations.

What carries the argument

The load-bearing mechanism is a directed spatial-diffusion graph $G=(V,E)$ over stations, with edge attributes carrying distance, angle, source wind speed, source wind direction, and an advection coefficient. The encoder runs TransformerConv message passing on this graph at each history time step, appends the resulting spatial representation to each station's features and PM2.5 history, and feeds the concatenation to a GRU. In the decoder, Luong attention computes a weighted combination of encoder hidden states before an MLP converts the combined state into a forecast. The graph is what encodes the physics of wind-driven transport, and the attention is what lets the decoder re-read history rather than relying on a single context vector.

What would settle it

Pick a test period in the Bihar or China data where the recorded wind components at a source station reverse direction between the history window and the forecast window, and compare AGNN_GRU's errors at downwind stations against a variant that reruns TransformerConv using the observed forecast-window meteorology; if the variant is materially better, the decoder's no-meteorology assumption is the limiting factor.

Watch

Extended reading notes

Core claim

The paper's core discovery is that spatial diffusion in PM2.5 can be represented as directed graph edges whose attributes include wind speed and wind direction at the source, distance, angle, and an advection coefficient, and that a TransformerConv message-passing step over this graph, combined with GRU temporal modeling and Luong attention in the decoder, yields a general spatio-temporal forecasting architecture. The paper asserts this architecture generalizes the ablation baselines (plain GRU, binary-edge GC_GRU, weighted-edge WGC_GRU, and GNN_GRU without attention) and reports that AGNN_GRU attains the best or near-best normalized loss, RMSE, and MAE, and the top Spearman rank correlation, CSI, and POD across 12-hour and 24-hour forecasts on both datasets.

Load-bearing premise

The model assumes that the spatial-diffusion patterns learned from historical wind and weather remain adequate during the forecast window, because the decoder receives no meteorological or graph input for future hours.

Editorial extensions

If this is right

  • Forecasts can be issued without requiring a separate weather forecast for the future window, removing a practical obstacle in diffusion-aware PM2.5 forecasting.
  • The reported results imply the model is better at preserving station-to-station pollution ordering (Spearman rank correlation) and at flagging haze-threshold crossings (CSI and POD) than the compared variants, which is the behavior an early-warning system needs.
  • The architecture's components are modular enough that GRU, GC_GRU, WGC_GRU, and GNN_GRU are special cases, so the ablation tables serve as a direct test of each component's contribution.
  • On the four-year China dataset the model reports consistently best statistics, suggesting the approach transfers across geographies and multi-year seasonality, not just the one-year Bihar deployment.

Reading between the lines

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

  • Editorial extension: If wind conditions shift sharply after the history window, the decoder's frozen graph may lag; a cheap upgrade would continue TransformerConv with last-known edge attributes or a learned persistence of wind fields, and the error difference would quantify how much the graph omission costs.
  • Editorial extension: The Bihar graph uses a 5 km distance threshold for computational reasons, yet the paper cites evidence of transport over hundreds of kilometres; a multi-scale or coarser auxiliary graph would test whether long-range advection is being captured at all.
  • Editorial extension: Because the paper says the architecture reduces to each baseline by removing components, one could reproduce the entire comparison by toggling attention, edge weighting, and graph direction on one code path; this would make the generalization claim directly checkable.
  • Editorial extension: If Luong attention is the main driver of long-horizon gains, then Spearman rank correlation should rise monotonically with history length only when attention is present; that comparison would isolate the mechanism.
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 / 5 minor

Summary. The paper proposes AGNN_GRU, a sequence-to-sequence architecture for PM2.5 forecasting in which an encoder combines TransformerConv graph convolutions with a GRU, and a decoder uses a GRU with Luong attention. The model is evaluated on a newly collected dense sensor network dataset from Bihar, India, and on the publicly available KnowAir dataset from China, under 12-hour and 24-hour forecast settings. The central claim is that the architecture captures spatial diffusion and temporal dependencies better than the compared baselines, and that it can be viewed as a generalization of several existing spatio-temporal forecasting models.

Significance. If the central claims were fully substantiated, the paper would offer a usefully simple and reproducible spatio-temporal forecasting model, and the newly collected Bihar dataset would be a valuable community resource. The paper ships public code and evaluates on a second, widely used dataset, which are genuine strengths. However, the current experimental evidence is limited to an ablation study against variants of the proposed model; no published spatio-temporal forecasting method is run on the same data. The significance of the empirical contribution is therefore currently smaller than the abstract claims, and the manuscript needs additional external comparisons and more careful per-setting reporting before the claims can be accepted.

major comments (4)
  1. [Section 4.2, Tables 2 and 3] The evaluation is only an ablation study: GRU, GC_GRU, WGC_GRU, and GNN_GRU are constructed by removing components from AGNN_GRU. No published spatio-temporal forecasting model, such as PM2.5-GNN [29], Airformer [16], or STGCN-type models, is run on either dataset. Consequently, the statement in Section 4.4.1 that 'our proposed model outperforms the baselines' supports only a component-contribution claim, not a claim of superiority over existing methods. Please add at least one published baseline trained under the same assumption that future meteorological covariates are unavailable, or explicitly justify why the omission is necessary, and temper the abstract and conclusions accordingly.
  2. [Section 4.4.2, Table 2 (bottom)] In the Bihar 24-hour forecast setting, AGNN_GRU does not achieve the best normalized loss, RMSE, or MAE: GC_GRU reports 0.27±0.02, 40.71±2.07, and 34.32±2.10, while AGNN_GRU reports 0.29±0.01, 42.56±1.22, and 37.38±1.23. The text acknowledges this only in part, but the claim that the model 'outperforms the baselines in almost all metrics' is not supported in this setting. The advantage of AGNN_GRU rests on Spearman correlation and CSI/POD/FAR, so the per-setting claims should be stated accurately and the statistical significance of the differences, given the reported standard deviations, should be discussed.
  3. [Section 4.3.1] The haze threshold used to binarize predictions for CSI, POD, and FAR is defined as the average PM2.5 concentration across all locations for the entire year, and the test period is part of that year. This makes the threshold a test-informed quantity. Please compute the threshold from the training and validation sets only, or demonstrate that the CSI/POD/FAR conclusions are insensitive to the threshold choice.
  4. [Abstract and Section 2.3] The claim that the model 'can also be seen as a generalization of various existing models' is asserted without a precise statement of which models are special cases and under what conditions. Because this claim appears in the abstract and motivates the architecture, please either formalize it (e.g., by showing that existing graph-based recurrent or encoder-decoder models are obtained by specific component choices) or remove it.
minor comments (5)
  1. [Section 2.3, Equation (5)] The notation in Equation (5) uses the same symbol for two different quantities: X_t^i appears on both sides of the bracket as if it were both the node attributes and the location/time attributes. Please use distinct symbols, as in the problem definition in Section 2.2.
  2. [Section 4.3.1, Equation (19)] Equation (19) writes the Pearson correlation formula, not Spearman's rank correlation. Spearman's rho requires first replacing the values by their ranks. Please correct the formula or clarify that the reported values are obtained by applying this formula to ranks.
  3. [Section 4.2] The statement that the ablated baselines 'represent special cases of several state-of-the-art models' is not demonstrated. Please name the specific models and explain the correspondence, or soften the wording.
  4. [Section 2.1] There is a typo in 'Total Precipiation' which should read 'Total Precipitation'. Please also proofread for similar minor grammatical issues, such as 'the model architecture to also differ significantly' in the Introduction.
  5. [Section 2.3] The decoder does not use a graph or GNN because future meteorological data are unavailable. This is a reasonable design choice, but it is also an important limitation: the spatial-diffusion patterns are fixed from historical data and cannot adapt to wind-field changes during the forecast horizon. Please state this limitation explicitly in the conclusions.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found: AGNN_GRU is specified by explicit equations, trained on held-out history windows, and evaluated on temporally separated test splits; the ablation-only baselines and test-informed haze threshold are benchmarking concerns, not derivation circularity.

full rationale

The derivation chain is self-contained. AGNN_GRU is defined by explicit equations (Eqs. 3-18), trained with MSE on observed history windows, and evaluated on temporally held-out validation and test splits shown in Table 1. The graph construction, edge attributes, and loss all use history features; no model parameter is fitted to the test labels. The Section 4.2 statement that the baselines are special cases of the proposed architecture is a structural claim and, if anything, makes the comparison an ablation rather than a derivation from the target result. The only test-informed quantity is the haze threshold in Section 4.3.1: 'which is the average PM2.5 concentration across all locations for the entire year.' Because the entire year includes the test partition, the CSI/POD/FAR event definitions are not fully independent of the labels being predicted; this is a mild evaluation-leakage concern and weakens the 'outperforms in almost all metrics' claim in the 24-hour Bihar row of Table 2, where the only wins are Spearman R, CSI, and POD. It does not, however, make the forecast values equal to, or forced by, the inputs or by any self-citation. Likewise, the absence of published external models among the baselines, and the omission of PM2.5-GNN due to differing problem settings, are missing-comparison limitations rather than circular reasoning. No load-bearing self-citations or uniqueness imports appear in the reference list.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The model's performance depends on user-chosen graph and evaluation thresholds, plus domain assumptions about how wind-driven diffusion is represented and whether ignoring future meteorology is safe. No fundamentally new physical entities are introduced.

free parameters (4)
  • Distance threshold for graph edges = 5 km (Bihar), 300 km (China)
    Determines which station pairs are connected in the diffusion graph; chosen for computational feasibility, not from physical considerations (Section 4.4.2).
  • Haze threshold for CSI/POD/FAR = 100 µg/m3 (Bihar), 75 µg/m3 (China)
    Set to the average PM2.5 concentration over the full year, which includes test data; affects the binary classification metrics (Section 4.3.1).
  • History and forecast lengths = H=24/48h, F=12/24h
    Experimental design choices against which all results are reported (Table 1).
  • Learning rate, regularization, embedding dimension = 1e-3, 3e-3, 8
    Selected by Bayesian search on the validation set (Section 4.3.2).
assumptions (3)
  • domain assumption Wind speed, wind direction, and advection coefficient as edge attributes capture inter-station PM2.5 transport.
    The diffusion graph is built on these attributes, and the model's spatial-diffusion claim rests on this representation (Section 2.3.1).
  • domain assumption Spatial dependencies learned from history are sufficient for forecasting without future meteorological inputs.
    The decoder uses no GNN or weather data, assuming transport patterns persist over the forecast window (Section 2.3).
  • standard math TransformerConv, GRU, and Luong attention behave as specified in their source papers.
    These standard building blocks are adopted without re-derivation (Equations 3-18).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spatio-Temporal Forecasting of PM2.5 via Spatial-Diffusion guided Encoder-Decoder Architecture." pith.science (2026). https://pith.science/paper/3VIQ6P7C

@misc{pith2026241213935,
  author       = {Pith},
  title        = {Pith review of: Spatio-Temporal Forecasting of PM2.5 via Spatial-Diffusion guided Encoder-Decoder Architecture},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3VIQ6P7C}},
  note         = {Machine review of arXiv:2412.13935}
}
abstract

In many problem settings that require spatio-temporal forecasting, the values in the time-series not only exhibit spatio-temporal correlations but are also influenced by spatial diffusion across locations. One such example is forecasting the concentration of fine particulate matter (PM2.5) in the atmosphere which is influenced by many complex factors, the most important ones being diffusion due to meteorological factors as well as transport across vast distances over a period of time. We present a novel Spatio-Temporal Graph Neural Network architecture, that specifically captures these dependencies to forecast the PM2.5 concentration. Our model is based on an encoder-decoder architecture where the encoder and decoder parts leverage gated recurrent units (GRU) augmented with a graph neural network (TransformerConv) to account for spatial diffusion. Our model can also be seen as a generalization of various existing models for time-series or spatio-temporal forecasting. We demonstrate the model's effectiveness on two real-world PM2.5 datasets: (1) data collected by us using a recently deployed network of low-cost PM$_{2.5}$ sensors from 511 locations spanning the entirety of the Indian state of Bihar over a period of one year, and (2) another publicly available dataset that covers severely polluted regions from China for a period of 4 years. Our experimental results show our model's impressive ability to account for both spatial as well as temporal dependencies precisely.

Figures

Figures reproduced from arXiv: 2412.13935 by the authors.

Figure 1
Figure 1. Architecture of the proposed AGNN_GRU model [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Distribution of AQM stations across the entire state [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Spatial Distribution of mean PM2.5 concentration for Bihar across seasons. From Left to Right: JJAS, ON, DJF, MA Model Loss RMSE MAE Spearman R CSI (%) POD (%) FAR (%) AGNN_GRU 0.22 ± 0.01 34.04 ± 1.30 29.77 ± 1.08 0.77 ± 0.01 48.27 ± 0.62 76.81 ± 2.50 43.42 ± 1.92 GNN_GRU 0.24 ± 0.02 35.22 ± 1.93 30.50 ± 1.94 0.73 ± 0.01 47.63 ± 0.56 72.56 ± 3.42 41.73 ± 2.67 WGC_GRU 0.24 ± 0.02 35.37 ± 1.88 30.66 ± 1.75 0.74 ± 0.0… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Comparison of true (blue color) and predicted values (orange color) across different locations for Bihar dataset for [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 19 canonical work pages

  1. [29]

    Shuo Wang, Yanran Li, Jiang Zhang, Qingye Meng, Lingwei Meng, and Fei Gao

  2. [16]

    Yuxuan Liang, Yutong Xia, Songyu Ke, Yiwei Wang, Qingsong Wen, Junbo Zhang, Yu Zheng, and Roger Zimmermann. 2023. Airformer: Predicting nationwide air quality in china with transformers. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 37. 14329–14337

  3. [1]

    Lukas Biewald. 2020. Experiment Tracking with Weights and Biases. https: //www.wandb.com/ Software available from wandb.com

  4. [2]

    Stef Buuren and Catharina Groothuis-Oudshoorn. 2011. MICE: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software 45 (12 2011). https://doi.org/10.18637/jss.v045.i03

  5. [3]

    Ling Chen, Jiahui Xu, Binqing Wu, and Jianlong Huang. 2023. Group-aware graph neural network for nationwide city air quality forecasting. ACM Transactions on Knowledge Discovery from Data 18, 3 (2023), 1–20

  6. [4]

    Quanchao Chen, Ruyan Ding, Xinyue Mo, Huan Li, Linxuan Xie, and Jiayu Yang. 2024. An adaptive adjacency matrix-based graph convolutional recurrent network for air quality prediction. Scientific Reports 14, 1 (2024), 4408

  7. [5]

    Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, and Yoshua Bengio. 2014. Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling. arXiv:1412.3555 [cs.NE] https://arxiv.org/abs/1412.3555

  8. [6]

    Xi Gao and Weide Li. 2021. A graph-based LSTM model for PM2.5 forecasting. Atmospheric Pollution Research 12, 9 (2021), 101150. https://doi.org/10.1016/j.apr. 2021.101150

Show all 38 references
  1. [7]

    Jindong Han, Hao Liu, Hengshu Zhu, Hui Xiong, and Dejing Dou. 2021. Joint air quality and weather prediction based on multi-adversarial spatiotemporal networks. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 35. 4081–4089

  2. [8]

    Jindong Han, Weijia Zhang, Hao Liu, and Hui Xiong. 2023. Machine learning for urban air quality analytics: A survey. arXiv preprint arXiv:2310.09620 (2023)

  3. [9]

    Junfeng Hu, Yuxuan Liang, Zhencheng Fan, Hongyang Chen, Yu Zheng, and Roger Zimmermann. 2023. Graph Neural Processes for Spatio-Temporal Extrapo- lation. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 752–763

  4. [10]

    Jianlin Hu, Yungang Wang, Qi Ying, and Hongliang Zhang. 2014. Spatial and temporal variability of PM2.5 and PM10 over the North China Plain and the Yangtze River Delta, China. Atmospheric Environment 95 (2014), 598–609. https: //doi.org/10.1016/j.atmosenv.2014.07.019

  5. [11]

    IQAir. 2021. World Air Quality Report 2020. https://www.iqair.com/world-most- polluted-cities/world-air-quality-report-2020-en.pdf Accessed: 2024-07-29

  6. [12]

    Guangyin Jin, Yuxuan Liang, Yuchen Fang, Zezhi Shao, Jincai Huang, Junbo Zhang, and Yu Zheng. 2023. Spatio-temporal graph neural networks for predictive learning in urban computing: A survey. IEEE Transactions on Knowledge and Data Engineering (2023)

  7. [13]

    Yong-been Kim, Seung-Bu Park, Sangchul Lee, and Young-Kwon Park. 2023. Comparison of PM2. 5 prediction performance of the three deep learning models: A case study of Seoul, Daejeon, and Busan. Journal of Industrial and Engineering Chemistry 120 (2023), 159–169

  8. [14]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. 2017. Semi-Supervised Classification with Graph Convolutional Networks. arXiv:1609.02907 [cs.LG] https://arxiv.org/abs/ 1609.02907

  9. [15]

    Yuxuan Liang, Haomin Wen, Yuqi Nie, Yushan Jiang, Ming Jin, Dongjin Song, Shirui Pan, and Qingsong Wen. 2024. Foundation models for time series analysis: A tutorial and survey. arXiv preprint arXiv:2403.14735 (2024)

  10. [17]

    Bryan Lim and Stefan Zohren. 2021. Time-series forecasting with deep learning: a survey. Philosophical Transactions of the Royal Society A 379, 2194 (2021), 20200209

  11. [18]

    Yachuan Liu, Jiaqi Ma, Paramveer Dhillon, and Qiaozhu Mei. 2024. PM2. 5 forecasting under distribution shift: A graph learning approach. AI Open 5 (2024), 23–29

  12. [19]

    Zhenyu Liu, Zhengtong Zhu, Jing Gao, and Cheng Xu. 2021. Forecast methods for time series data: a survey. Ieee Access 9 (2021), 91896–91912

  13. [20]

    Minh-Thang Luong, Hieu Pham, and Christopher D. Manning. 2015. Effective Approaches to Attention-based Neural Machine Translation. arXiv:1508.04025 [cs.CL] https://arxiv.org/abs/1508.04025

  14. [21]

    John A Miller, Mohammed Aldosari, Farah Saeed, Nasid Habib Barna, Subas Rana, I Budak Arpinar, and Ninghao Liu. 2024. A survey of deep learning and foundation models for time series forecasting. arXiv preprint arXiv:2401.13912 (2024)

  15. [22]

    Sara Mirzavand Borujeni, Leila Arras, Vignesh Srinivasan, and Wojciech Samek

  16. [23]

    Hamilton, Jan Eric Lenssen, Gaurav Rattan, and Martin Grohe

    Christopher Morris, Martin Ritzert, Matthias Fey, William L. Hamilton, Jan Eric Lenssen, Gaurav Rattan, and Martin Grohe. 2021. Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks. arXiv:1810.02244 [cs.LG] https: //arxiv.org/abs/1810.02244

  17. [24]

    Lisanne Petry, Thomas Meiers, David Reuschenberg, Sara Mirzavand Borujeni, Jost Arndt, Luise Odenthal, Thilo Erbertseder, Hannes Taubenböck, Inken Müller, Elena Kalusche, et al. 2021. Design and results of an AI-based forecasting of air pollutants for smart cities. ISPRS Annal...

  18. [25]

    Yunsheng Shi, Zhengjie Huang, Shikun Feng, Hui Zhong, Wenjin Wang, and Yu Sun. 2021. Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification. arXiv:2009.03509 [cs.LG] https://arxiv.org/abs/ 2009.03509

  19. [26]

    Avikal Somvanshi, Sharanjeet Kaur, and Anumita Roychowdhury. 2024. End of winter report 2023-24: Spread and scale of air pollution crisis in India. (04 2024)

  20. [27]

    Mengfan Teng, Siwei Li, Jia Xing, Chunying Fan, Jie Yang, Shuo Wang, Ge Song, Yu Ding, Jiaxin Dong, and Shansi Wang. 2023. 72-hour real-time forecasting of ambient PM2.5 by hybrid graph deep neural network with aggregated neighbor- hood spatiotemporal information. Environment ...

  21. [28]

    Mengfan Teng, Siwei Li, Jie Yang, Jiarui Chen, Chunying Fan, and Yu Ding. 2024. A new hybrid deep neural network for multiple sites PM2.5 forecasting. Journal of Cleaner Production 473 (2024), 143542. https://doi.org/10.1016/j.jclepro.2024. 143542

  22. [30]

    Haomin Wen, Youfang Lin, Yutong Xia, Huaiyu Wan, Qingsong Wen, Roger Zimmermann, and Yuxuan Liang. 2023. Diffstg: Probabilistic spatio-temporal graph forecasting with denoising diffusion models. InProceedings of the 31st ACM International Conference on Advances in Geographic I...

  23. [31]

    Yu-Fei Xing, Yue-Hua Xu, Min-Hua Shi, and Yi-Xin Lian. 2016. The impact of PM2.5 on the human respiratory system. Journal of Thoracic Disease 8, 1 (2016). https://jtd.amegroups.org/article/view/6353

  24. [32]

    Yiyuan Yang, Ming Jin, Haomin Wen, Chaoli Zhang, Yuxuan Liang, Lintao Ma, Yi Wang, Chenghao Liu, Bin Yang, Zenglin Xu, et al. 2024. A survey on diffusion models for time series and spatio-temporal data. arXiv preprint arXiv:2404.18886 (2024)

  25. [33]

    Chin-Chia Michael Yeh, Xin Dai, Huiyuan Chen, Yan Zheng, Yujie Fan, Audrey Der, Vivian Lai, Zhongfang Zhuang, Junpeng Wang, Liang Wang, et al . 2023. Toward a foundation model for time series data. In Proceedings of the 32nd ACM International Conference on Information and Know...

  26. [34]

    Lingxiao Zhao, Zhiyang Li, and Leilei Qu. 2022. Forecasting of Beijing PM2.5 with a hybrid ARIMA model based on integrated AIC and improved GS fixed- order methods and seasonal decomposition. Heliyon 8, 12 (2022), e12239. https: //doi.org/10.1016/j.heliyon.2022.e12239

  27. [35]

    Hongye Zhou, Feng Zhang, Zhenhong Du, and Renyi Liu. 2021. Forecasting PM2. 5 using hybrid graph convolution-based model considering dynamic wind-field to offer the benefit of spatial interpretability. Environmental Pollution 273 (2021), 116473

  28. [36]

    Shiyun Zhou, Wei Wang, Long Zhu, Qi Qiao, and Yulin Kang. 2024. Deep-learning architecture for PM2. 5 concentration prediction: A review.Environmental Science and Ecotechnology (2024), 100400

  29. [2020]

    In Proceedings of the 28th International Conference on Advances in Geographic Information Systems (SIGSPATIAL ’20)

    PM2.5-GNN: A Domain Knowledge Enhanced Graph Neural Network For PM2.5 Forecasting. In Proceedings of the 28th International Conference on Advances in Geographic Information Systems (SIGSPATIAL ’20) . ACM. https: //doi.org/10.1145/3397536.3422208

  30. [2023]

    Scientific Reports 13, 1 (2023), 9940

    Explainable sequence-to-sequence GRU neural network for pollution forecasting. Scientific Reports 13, 1 (2023), 9940

Pith tools

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