Public Bike Trip Purpose Inference Using Point-of-Interest Data

Public bike-sharing is eco-friendly, connects excellently with other transportation modes, and provides a means of mobility that highly suitable in the current era climate change. This study proposes methodology for inferring bike trip purpose based on bike-share point-of-interest (POI) data. Because involves decision-making, its inference necessitates an understanding spatiotemporal complexity human activities. Thus, features affecting trips were selected from data, land uses at origin destination extracted POI During type embedding, data augmented considering geographical distance between POIs number rentals each station. We further developed ground truth construction method temporal mobile The model was built using machine learning applied to experiments involving stations Seocho-gu, Seoul, Korea. experimental results revealed optimal performance achieved use decision tree algorithms, as demonstrated by 78.95% overall accuracy 66.43% F1-score. proposed contributes better causes movement within cities.