Predictive Model Analysis For School Bus Services In Jakarta Using A Machine Learning Approach

Abstract

This research aims to predict the types of school buses in Jakarta using machine learning methods. Data from 2017 to 2019 includes the number of passengers, number of schools, and bus types. Exploratory data analysis identified patterns and trends, with feature engineering generating three main variables. Seven machine learning models were tested, including SVM, Logistic Regression, KNN, Gaussian Naive Bayes, Decision Tree, AdaBoost, and Gradient Boosting, with a focus on f1-score to handle data imbalance. The evaluation shows that Gradient Boosting has the best performance with the highest accuracy, precision, recall, and f1-score. The results provide insights into the factors that influence school bus types and offer an effective predictive model to support decision-making in school transportation management in Jakarta. Gradient Boosting proved to be the most reliable in predicting school bus types, providing a basis for strategies to improve the safety and efficiency of school transportation.