| 研究実績の概要 |
In 2015 fiscal year, our achievements of this project can be summarize as follows: (1) We propose a novel model called CityMomentum as a predicting-by-clustering framework for predicting short-term crowd behavior at a citywide level. (2) We develop a deep model of Stack Denoise Autoencoder to learn hierarchical feature representation of human mobility. Then these features are used for efficient prediction of traffic accident risk level. Our model can simulate corresponding traffic accident risk map with the given real-time input of human mobility.
Our research results were published in the eminent publications for computer science including UbiComp 2015 and AAAI 2016. Our research results on predicting human crowd behavior received honorable mention award in UbiComp 2015.
|
