University of the Philippines
In this paper, the authors characterized traffic density modeled from coarse data by using data signatures to effectively and efficiently represent traffic flow behavior. Using the 2006 North Luzon EXpressway North Bound (NLEX NB) BaLintawaK (Blk), Bocaue (Boc), Meycauayan (Mcy), and Marilao (Mrl) segments' hourly traffic volume and time mean speed data sets provided by the National Center for Transportation Studies (NCTS), they generated hourly traffic density data set. Each point in the data was represented by a 4D data signature where cluster models and 2D visualizations were formulated and varying traffic density behaviors were identified, i.e., high and low traffic congestions, outliers, etc. Best-fit curves, confidence bands and ellipses were generated in the visualizations for additional cluster information.