Prediction of Preferences for Public Transport Car Types using Machine Learning
Agus Pratondo, Aprianti Putri Sujana
Abstract
Understanding customer preferences for various modes, such as angkot (public minivans) and buses, is crucial given the importance of public transportation for urban mobility. The Random Forest technique is used in a prediction analysis to determine consumer preferences between angkot and buses. Evaluation of the model's performance in terms of Precision, Recall, F1-Score, and Accuracy was the study's main objective. Its crucial features that influence user choice were also noted. As part of the research, detailed user preferences, demographic data, and travel behaviors were gathered. An ensemble learning technique known for its effectiveness in classification tasks, Random Forest, was used to estimate consumer transportation preferences. The investigation yielded encouraging results, with the prediction model obtaining high levels of accuracy. The model's outstanding Precision of 0.979 indicates the precision of the model's positive predictions (preferring angkot). In addition, the model achieved a Recall of 1.000, indicating its ability to correctly identify positive events (correctly predicting angkot preferences). The F1-Score was calculated to be 0.989, which represents the model's overall performance. In addition, the study effectively discovered a number of critical characteristics that greatly influenced the classification of angkot and buses. These crucial factors offer valuable insights into the underlying reasons behind users' preferences and provide transportation authorities with guidance to enhance public transit planning and service optimization.