An architecture for emergency event prediction using LSTM recurrent neural networks

作者：

Highlights：

• An architecture for emergency event prediction is proposed.

• Binary classification and regression models are developed.

• LSTM recurrent neural networks are adopted.

• The proposed models overwhelmed time series forecasting and machine learning.

• The assumption on spatial dependency was evaluated.

摘要

•An architecture for emergency event prediction is proposed.•Binary classification and regression models are developed.•LSTM recurrent neural networks are adopted.•The proposed models overwhelmed time series forecasting and machine learning.•The assumption on spatial dependency was evaluated.

论文关键词：Emergency events,Emergency prediction system,Recurrent neural network,Long short-term memory

论文评审过程：Received 24 July 2017, Revised 18 December 2017, Accepted 20 December 2017, Available online 20 December 2017, Version of Record 30 December 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.12.037