Scholarlink

<p>Air pollution is one of the most serious environmental risks to public health worldwide. Therefore, the accurate prediction of daily air quality levels holds significant importance in assessing and managing urban air pollution. This study has exploited the ensemble intelligence of tree-based models in the first-ever attempt to predict the Air Quality Index (AQI) for Quetta City of Pakistan into three categories: good, moderate, and unhealthy levels. Meteorological and pollutant parameters are both considered for categorization of AQI levels. The class imbalance of AQI data has been addressed by employing SMOTE oversampling, whereas multiple tree-based ensembles, namely Random Forest (RF), Balanced RF, Shallow RF, and Extra Trees, have been optimized to predict AQI of Quetta city, known as driest city of Pakistan. The SRF in combination with SMOTE demonstrates higher performance, making it the most promising model for reliable AQI assessment (Accuracy: 98.7%, Precision: 98.8%, Recall: 98.7%, F1-score: 98.5%). The study also identifies particulate matter, temperature, and humidity as the most influential predictors for the categorization of AQI levels, thus supporting environmental management and risk mitigation in a dry and polluted urban environment of Quetta.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AQI prediction for dry climate: A case study of Quetta City using ensemble learning

  • 期刊论文

  • 发表日期：

    2025年10月13日

查看全文
  • Ubaid Ullah,
  • Adnan Idris,
  • Raja Asif Wagan

摘要

Air pollution is one of the most serious environmental risks to public health worldwide. Therefore, the accurate prediction of daily air quality levels holds significant importance in assessing and managing urban air pollution. This study has exploited the ensemble intelligence of tree-based models in the first-ever attempt to predict the Air Quality Index (AQI) for Quetta City of Pakistan into three categories: good, moderate, and unhealthy levels. Meteorological and pollutant parameters are both considered for categorization of AQI levels. The class imbalance of AQI data has been addressed by employing SMOTE oversampling, whereas multiple tree-based ensembles, namely Random Forest (RF), Balanced RF, Shallow RF, and Extra Trees, have been optimized to predict AQI of Quetta city, known as driest city of Pakistan. The SRF in combination with SMOTE demonstrates higher performance, making it the most promising model for reliable AQI assessment (Accuracy: 98.7%, Precision: 98.8%, Recall: 98.7%, F1-score: 98.5%). The study also identifies particulate matter, temperature, and humidity as the most influential predictors for the categorization of AQI levels, thus supporting environmental management and risk mitigation in a dry and polluted urban environment of Quetta.