EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

QualityNet: Complexity‐Aware Temporal–Spatial Air Quality Categorization via Deep Tabular Attention Networks on Multisensor Urban Data

Kishor Kumar Reddy C., Hafsa Ihteshamuddin Ahmed, Anisha P. R., Surbhi B. Khan, Asma Alshuhail, Amina Salhi

Complexity · 2026-01 · 2026 (1)

DOI: 10.1155/cplx/9155881 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Complexity, online erschienen am 30.07.2026. Verfasst von Kishor Kumar Reddy C. und 5 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Princess Nourah Bint Abdulrahman University, King Faisal University.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Air pollution is a serious threat to public health and urban sustainability, posing a threat to our planet’s ecological balance and human well‐being. Prolonged exposure to air pollution is linked to cardiovascular and respiratory diseases, cancers, and premature deaths. Urban locations, in particular, serve as a central hub for pollutant emissions from transportation vehicles, industrial units, and household operations, along with various meteorological factors.

Methods

This research proposes temporal–spatial (implicit) air quality categorization based on deep tabular attention networks, the QualityNet model, on multisensor urban air pollution data from the UCI Air Quality dataset, which contains continuous pollutant concentrations. These are later mapped into discrete air quality categories to facilitate a classification structure that has the capability to detect subtle nonlinearities in relationships between environmental variables and multiple gas sensor readings. This research addresses constraints such as class imbalance through weighted loss functions, feature scaling, and solid train–test splits to ensure model generalizability.

Results

Experimental results prove that the QualityNet classifier learns effective representative features from high‐dimensional sensor data with an accuracy of 95.78%, precision of 95.8%, recall of 95.78%, F1‐score of 95.78%, Cohen kappa score of 94.18%, and an AUC‐ROC score of 99.46%. The attentive mechanisms of the model are useful for revealing the most impactful sensor inputs, making environmental monitoring explainable and interpretable.

Conclusion

This paper emphasizes the capability of deep tabular learning architectures for robust, dynamic, and interpretable air quality classification in urban monitoring systems. It efficiently classifies air quality levels into multiple categories, achieving high evaluation metrics such as accuracy: 95.78% and AUC‐ROC: 0.9946. This provides a benchmark for future work in environmental and urban health monitoring.

Förderung

  • Princess Nourah Bint Abdulrahman University
  • King Faisal University

Zitieren

APA

C., K. K. R., Ahmed, H. I., R., A. P., Khan, S. B., Alshuhail, A., & Salhi, A. (2026). QualityNet: Complexity‐Aware Temporal–Spatial Air Quality Categorization via Deep Tabular Attention Networks on Multisensor Urban Data. Complexity, 2026(1). https://doi.org/10.1155/cplx/9155881

Vancouver

C. KKR, Ahmed HI, R. AP, Khan SB, Alshuhail A, Salhi A. QualityNet: Complexity‐Aware Temporal–Spatial Air Quality Categorization via Deep Tabular Attention Networks on Multisensor Urban Data. Complexity. 2026;2026(1). doi:10.1155/cplx/9155881

BibTeX
@article{C2026,
  author = {C., Kishor Kumar Reddy and Ahmed, Hafsa Ihteshamuddin and R., Anisha P. and Khan, Surbhi B. and Alshuhail, Asma and Salhi, Amina},
  title = {QualityNet: Complexity‐Aware Temporal–Spatial Air Quality Categorization via Deep Tabular Attention Networks on Multisensor Urban Data},
  journal = {Complexity},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/cplx/9155881}
}
Kostenlos starten →