Utilizing Real-World Transportation Data for Accurate Traffic Prediction

Abstract

Abstract-For the first time, real-time high-fidelity spatiotemporal data on transportation networks of major cities have become available. This gold mine of data can be utilized to learn about traffic behavior at different times and locations, potentially resulting in major savings in time and fuel, the two important commodities of the 21st century. As a first step towards the utilization of this data, in this paper, we study the real-world data collected from Los Angeles County transportation network in order to incorporate the data's intrinsic behavior into a time-series mining technique to enhance its accuracy for traffic prediction. In particular, we utilized the spatiotemporal behaviors of rush hours and events to perform a more accurate prediction of both shortterm and long-term average speed on road-segments, even in the presence of infrequent events (e.g., accidents). Our result shows that taking historical rush-hour behavior we can improve the accuracy of traditional predictors by up to 67% and 78% in short-term and long-term predictions, respectively. Moreover, we can incorporate the impact of an accident to improve the prediction accuracy by up to 91%.

A review of urban computing for mobile phone traces

“…We have identified as suitable options PostgreSQL extended with PostGIS to store and manipulate the cell phone data and the spatial networks, and pgRouting to add geospatial routing capabilities to the database. Beyond the availability of numerous features that make it possible to conduct various analyses, these open source tools provide a great foundation upon which to implement extensions related to time dependency [34,10,9] and multimodality [7] for example.…”

Context-Aware Online Spatiotemporal Traffic Prediction

“…The baseline approaches that we compare against include the single base predictors using SVM and NB and ensemble learning techniques based weighted majority [11]- [13]. We also compare with our prior work [1], labeled by "PAN".…”

“…We use θ t ∈ Θ to denote the context information associated with the t-th incident where Θ is a D-dimensional space and D is the number of types of context used. Without loss of generality, we normalize the context space Θ to be [0, 1] D . For each incident, based on the context information, the system selects the prediction of one of the base predictors as the final traffic prediction, denoted by y t ∈ Y.…”

“…One challenge in predicting traffic is how much to rely on the historical data vs. the real-time data. Previous studies [1] [2] showed that depending on the situation one dataset may be more useful than the other but there is no holistic approach on when to switch from one dataset to the other for a more effective prediction. This becomes even more challenging when considering different causes for congestions.…”

“…The majority of traffic prediction techniques focused on predicting traffic in typical conditions (e.g., morning rush hours) [1], [3]- [5], and more recently in the presence of accidents (e.g., [3], [6]. Existing techniques are only applicable to predict one of the scenarios.…”