A Conceptual Framework for Bus Arrival Prediction Based on Spark Framework and Machine Learning Approaches

• Main Authors: Li Ye, Pree Thiengburanathum, Poon Thiengburanathum
• Format: Conference Proceeding
• Published: 2020
• Subjects: Arts and Humanities; Computer Science; Energy; Engineering; Medicine
• Online Access: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85085627062&origin=inward; http://cmuir.cmu.ac.th/jspui/handle/6653943832/70122
• Institution: Chiang Mai University

© 2020 IEEE. Recently, the bus system has been deployed as a new option for public transportation in Chiang Mai. However, the reason why the people hesitate to take the bus because they are not confident about accuracy of the bus schedule. Predicting the bus arriving time at a certain bus station is a challenging problem, due to concerns with real-Time data pre-processing, numerous data inputs, and the level of predictive accuracy. Previous studies that related to bus time arrival prediction have applied statistical and machine-learning methods. However, the time series problem is rarely considered in the previous machine-learning prediction methods. Moreover, some models only analyze small amounts of data, which leads to poor accuracy and poor speed of prediction. This paper reviewed previous five years studies related to the prediction of bus arrival time. The paper highlights the current research gaps and applications of bus arrival time prediction. We proposed the research framework, as well as the possible research trends and challenges.