ABSTRACT. This study is based on a machine learning algorithm known as
XGBoost. We used the XGBoost algorithm to forecast the capacity of Jakarta's
mass transit system. We used preprocessed raw data from the Jakarta Open Data
website for 2020-2021 as a training medium to achieve a mean absolute
percentage error of 69. However, after the model was fine-tuned, the MAPE was
significantly reduced by 28.99% to 49.97. The XGBoost algorithm effectively
detected patterns and trends in the data, which can be used to improve routes and
plan future studies by providing valuable insights. It is possible that additional
data points, such as holidays and weather conditions, will further enhance the
model's accuracy in future research. As a result of implementing XGBoost,
Jakarta's transportation system can optimize resource utilization and improve
customer service to improve passenger satisfaction. Future studies may benefit
from additional data points, such as holidays and weather conditions, to improve
XGBoost's efficiency.