This study proposes a novel approach to predicting travel mode choices using imbalance machine learning techniques. With the rapid urbanization and increasing complexity of transportation systems, understanding and predicting travel mode choices has become crucial for efficient transportation planning and management. The dataset used in this research includes various features such as cost, duration, and demographic attributes, which often exhibit imbalance across different travel modes. To address this challenge, we employ a semi-supervised learning approach that leverages both labeled and unlabeled data, enhancing model accuracy and robustness. Feature selection is performed using the K-Best method, identifying the most significant predictors of travel mode. The Support Vector Machine (SVM) model is then trained on these selected features, achieving notable improvements in classification metrics. The results demonstrate that our approach effectively handles data imbalance, providing accurate predictions and valuable insights into travel behavior. This study contributes to the field of transportation analytics by offering a robust methodology for predicting travel mode choices, which can aid policymakers and urban planners in making informed decisions.

Keywords:

Travel Mode Choice, Imbalance Machine Learning, Semi-Supervised Learning, Feature Selection, K-Best Method, Support Vector Machine, Transportation Planning, Urban Mobility, Data Imbalance, Predictive Modeling.