<![CDATA[One important component for the sustainability of living organisms is the availability of clean air. Air pollution has detrimental repercussions such as deteriorating air quality, which can have harmful impacts on the environment. As a result, modeling and monitoring air quality are essential steps in reducing air pollution, and the Air Quality Index allows for the review of this monitoring. This study's objective is to apply the Modified K-Nearest Neighbor method in classifying urban Air Quality Index data from various countries. The "World Air Quality Index by City and Coordinates" Kaggle website provided the data that were used. The data were then processed and tested using the Modified K-Nearest Neighbor model, this builds on the K-Nearest Neighbor technique. 16,695 data in total were divided into 80% training data (13,356 data) and 20% testing data (3,339 data) in order to conduct testing. Evaluation was performed by comparing the performance of the Modified K-Nearest Neighbor method with the traditional K-Nearest Neighbor method. The Modified K-Nearest Neighbor approach with K=1 produced an accuracy rate of 99.73% during testing, whereas the K-Nearest Neighbor method produced an accuracy rate of 99.64%. Using the K-Fold Cross Validation, K-Nearest Neighbor method perfrom highest mean score of 99,04% and Modified K-Nearest Neighbor perform highest mean score of 99,46%. ]]>
Copyrights © 2025
