カンファレンス (国際) CityOutlook: Early Crowd Dynamics Forecast towards Irregular Events Detection with Synthetically Unbiased Regression
Soto Anno (Tokyo Institute of Technology), Kota Tsubouchi, Masamichi Shimosaka (Tokyo Institute of Technology)
29th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM SIGSPATIAL 2021)
his study investigates crowd dynamics forecast one week in advance to detect irregular urban events, which plays an important role in risk-aware decision-making in urban regions such as congestion mitigation or crowd control for public safety. Although previous approaches have addressed crowd dynamics prediction, they have failed to deal with the scarcity of anomalous events, which results in a large model bias and could not quantify the number of visitors in anomalous crowd gathering. To efficiently handle this problem and provide an elaborate early forecast, we focus on the successive properties of importance weighting (IW) to penalize the anomalous data in terms of model bias; however, leveraging the concept of IW is challenging for two reasons: (1) Dividing dataset into normal and abnormal sets is difficult. (2) Adopting IW to the scarce data risks leading to the instability of the model. Motivated by these challenges, we propose CityOutlook, a novel forecasting model based on unbiased regression with importance-based synthetically resampling. To make IW applicable to our approach, we design an anomaly-aware data annotation scheme by utilizing the heterogeneous property of mobility data to determine the data anomaly. Furthermore, we develop an importance-based resampling algorithm to mitigate the model learning stability. We evaluate CityOutlook using the datasets of large-scale mobility and transit search logs. The experimental results show that CityOutlook outperforms the state-of-the-art models on crowd anomaly forecast, providing the same level accuracy in forecasting normal dynamics.