GSNet:Learning Spatial-Temporal Correlation from Geographical and Semantic Aspects for Traffic Accid
GSNet:Learning Spatial-Temporal Correlation from Geographical and Semantic Aspects for Traffic Accident Risk Forecasting(AAAI 2021)
提出问题:
导致交通事故因素很多(天气、时间、交通流量、POI等),以往的模型考虑不全面;
零值膨胀(zero-inflated)问题:零值过多会导致预测不准确;
为提高预测准确性:考虑区域各方面的区域相关性:.
原创
2021-05-25 16:49:16 ·
1123 阅读 ·
1 评论