<![CDATA[Improvement in air quality in urban areas like Central Jakarta is a big challenge due to high activities of transport, industry, and dense population. This study aims to predict PM2.5 concentrations by utilising the XGBoost algorithm based on temperature data as the main variable. The data was taken from Kemayoran, Central Jakarta, with an observation time span from 01 January 2017 to 12 February 2017. XGBoost was chosen due to the non-linear and complex nature of the data. Based on the results of the test, it shows that the model performance is far from improved, characterized by a high Mean Squared Error (MSE) value and a small R² score. These performance limitations are driven by the small amount of data and the absence of other supporting variables such as air humidity, wind speed, and rainfall. The high PM2.5 concentration was contributed by the research location in Kemayoran, one of the most densely populated areas with high industrial activity and fossil-fuelled transport. This study provides evidence to support the addition of supporting variables and the extension of the observation time span to enhance model accuracy. Therefore, the XGBoost algorithm can be used as a promising solution for air quality prediction in urban cities where air pollution has reached its peak.]]>
Copyrights © 2022
