Automated Transportation Mode Detection Using Smart Phone Applications via Machine Learning: Case Study Mega City of Tehran

Abstract

Through the past few decades, travel behaviors have become more complicated especially in mega cities such as Tehran, the capital of Iran. Decision makers require more accurate and comprehensive information to plan for city transportation. As opposed to traditional paper- based and telephone-based surveys, a new efficient and effective data collection method has been recently applied using information technology such as the global positioning system (GPS)-based data collection method which can track passengers’ trips. Having utilized this new method, the main aim of this study is to analyze the collected data in order to distinguish transportation modes used by passengers using a novel machine learning method called random forest. This model not only classifies transportation modes i.e., car, bus, and walking at a high accuracy of almost 96%, but also determines the most influential attributes in the process of classification based on two importance indices: mean decrease accuracy and Gini index. Results show that instant speed and accuracy of GPS track are the most influential attributes in the transportation modes classification. Transportation planners benefit a lot from this accurate and comprehensive travel behaviors data (used modes) for policy making.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 30 citations exist upstream.