Enhancing Transportation Mode Prediction from GPS Trajectories using Machine Learning and Feature Engineering

This manuscript proposes a comprehensive framework for the automated determination of travel modes based solely on GPS trajectories. To improve prediction accuracy, additional preprocessing features are introduced, including speed, acceleration, jerk, and bearing rate. Our approach employs various machine learning techniques, such as Random Forest, Multilayer Perceptron (MLP), AdaBoost Classifier, Decision Tree, Support Vector Machine (SVM), and K-Nearest Neighbors (KNN), to achieve notable classification results. Extensive evaluations demonstrate that our framework surpasses existing state-of-the-art algorithms for transport mode prediction. This investigation presents a promising approach for accurate and reliable prediction of travel modes, with potential applications in various real-world scenarios..