The goal of precipitation nowcasting is to predict the future rainfall
intensity in a local region over a relatively short period of time. Very few
previous studies have examined this crucial and challenging weather forecasting
problem from the machine learning perspective. In this paper, we formulate
precipitation nowcasting as a spatiotemporal sequence forecasting problem in
which both the input and the prediction target are spatiotemporal sequences. By
extending the fully connected LSTM (FC-LSTM) to have convolutional structures
in both the input-to-state and state-to-state transitions, we propose the
convolutional LSTM (ConvLSTM) and use it to build an end-to-end trainable model
for the precipitation nowcasting problem. Experiments show that our ConvLSTM
network captures spatiotemporal correlations better and consistently
outperforms FC-LSTM and the state-of-the-art operational ROVER algorithm for
precipitation nowcasting.
%0 Generic
%1 shi2015convolutional
%A Shi, Xingjian
%A Chen, Zhourong
%A Wang, Hao
%A Yeung, Dit-Yan
%A Wong, Wai-kin
%A Woo, Wang-chun
%D 2015
%K lstm prediction rain rainfall
%T Convolutional LSTM Network: A Machine Learning Approach for
Precipitation Nowcasting
%U http://arxiv.org/abs/1506.04214
%X The goal of precipitation nowcasting is to predict the future rainfall
intensity in a local region over a relatively short period of time. Very few
previous studies have examined this crucial and challenging weather forecasting
problem from the machine learning perspective. In this paper, we formulate
precipitation nowcasting as a spatiotemporal sequence forecasting problem in
which both the input and the prediction target are spatiotemporal sequences. By
extending the fully connected LSTM (FC-LSTM) to have convolutional structures
in both the input-to-state and state-to-state transitions, we propose the
convolutional LSTM (ConvLSTM) and use it to build an end-to-end trainable model
for the precipitation nowcasting problem. Experiments show that our ConvLSTM
network captures spatiotemporal correlations better and consistently
outperforms FC-LSTM and the state-of-the-art operational ROVER algorithm for
precipitation nowcasting.
@misc{shi2015convolutional,
abstract = {The goal of precipitation nowcasting is to predict the future rainfall
intensity in a local region over a relatively short period of time. Very few
previous studies have examined this crucial and challenging weather forecasting
problem from the machine learning perspective. In this paper, we formulate
precipitation nowcasting as a spatiotemporal sequence forecasting problem in
which both the input and the prediction target are spatiotemporal sequences. By
extending the fully connected LSTM (FC-LSTM) to have convolutional structures
in both the input-to-state and state-to-state transitions, we propose the
convolutional LSTM (ConvLSTM) and use it to build an end-to-end trainable model
for the precipitation nowcasting problem. Experiments show that our ConvLSTM
network captures spatiotemporal correlations better and consistently
outperforms FC-LSTM and the state-of-the-art operational ROVER algorithm for
precipitation nowcasting.},
added-at = {2017-11-13T21:33:04.000+0100},
author = {Shi, Xingjian and Chen, Zhourong and Wang, Hao and Yeung, Dit-Yan and Wong, Wai-kin and Woo, Wang-chun},
biburl = {https://www.bibsonomy.org/bibtex/2ae9fac5693c51a34cec5ebb9537f4cdd/lautenschlager},
interhash = {b6d0c65d2d62ed0824cd0bcc9ae5ddd4},
intrahash = {ae9fac5693c51a34cec5ebb9537f4cdd},
keywords = {lstm prediction rain rainfall},
note = {cite arxiv:1506.04214},
timestamp = {2017-11-13T21:33:04.000+0100},
title = {Convolutional LSTM Network: A Machine Learning Approach for
Precipitation Nowcasting},
url = {http://arxiv.org/abs/1506.04214},
year = 2015
}