Estimating the Missing Traffic Speeds via Continuous Conditional Random Fields

摘要

Recently, increasing interests have been emerging in the data driven intelligent transportation systems, some typical applications include flock pattern recognition, road network structure inference and route searching. However, almost all the applications suffer from the missing data problem. In this paper, we propose to adopt the Continuous Conditional Random Fields (CCRFs) model to estimate the missing historical traffic data. We exam the proposed method with a real traffic speed dataset, results show that it is superior to the comparison algorithms.