A unified framework for online trip destination prediction

Trip destination prediction is an area of increasing importance in many applications such as trip planning, autonomous driving and electric vehicles. Even though this problem could be naturally addressed online learning paradigm where data arriving a sequential fashion, the majority research has rather considered offline setting. In paper, we present unified framework for setting, which suitable both training prediction. For purpose, develop two clustering algorithms integrate them within models problem. We investigate different configurations on real-world dataset. demonstrate that entire yield consistent results compared to Finally, propose novel regret metric evaluating comparison its counterpart. This makes it possible relate source erroneous predictions either or model. Using metric, show proposed methods converge probability distribution resembling true underlying with lower than all baselines.