<![CDATA[Air quality is a critical indicator for public health and the environment. This study presents the first web‑based implementation for classifying the Air Pollutant Standard Index (ISPU) of Jakarta using the 2024 dataset from Satu Data Indonesia. The Gaussian Naive Bayes method was chosen for its efficiency and ability to handle continuous numerical data. Preprocessing steps included mean imputation, removal of “no data” and “very unhealthy” categories, and a random state 80:20 train‑test split. Evaluation results show 90.57% accuracy surpassing the KNN baseline of 86% with precision and recall F1‑scores for the “Unhealthy” category at 84.09% and 92.50%, respectively. A Flask‑based web application air quality prediction. These findings confirm the superiority of Gaussian Naive Bayes over KNN in handling data imbalance, while providing an accurate, accessible environmental monitoring tool. Contributions of this research include (1) the first deployment of ISPU Jakarta 2024 in a web system, and (2) a measured performance comparison between GNB and KNN.]]>
Copyrights © 2025
