Machine learning has been widely used in some fields, but has yet to really catch on in transport planning practice. However, research to date shows high promise for machine learning to complement existing approaches to understanding and predicting travel behavior. This two-part webinar will convey a broad introduction to the underlying concepts and methods, demystifying the confusing terminology and concepts within the context of transport planning applications.
Rick Donnelly (WSP) will present this overview, and will be joined by colleagues Kyle Ward (Caliper Corporation) and Mausam Duggal (WSP) to show real-world examples of how machine learning has been applied in income imputation, mode choice, and analysis of truck GPS tracking data. The examples will walk the audience through how the data were prepared, run through training using different machine learning packages, how the results can be used for prediction, and observations on the process.
Background in linear algebra, probability and statistics will be helpful but not required for this overview.