Deep learning methods for spatio-temporal correlation patterns
This document has been machine translated.
Deep learning methods (CRNN, LGBM) can be applied to event data to learn and predict spatio-temporal correlation patterns.
Convolutional Recurrent Neural Network (CRNN), which integrates spatial pattern learning by Convolutional Neural Network (CNN) and temporal pattern learning by Long Short Term Memory (LSTM), is used for environmental data such as weather and atmosphere. Neural Network (CRNN) method, which integrates spatial pattern learning by CNN and temporal pattern learning by Long Short Term Memory (LSTM), to learn and predict spatio-temporal correlation patterns.
In addition, by combining CRNN with LightGBM, it is possible to predict changes in environmental data (at present, this service is limited to photochemical oxidant prediction cases).
For more information about CRNN, please refer to the following paper.
- Zhao, P. and Zettsu, K.: Convolution Recurrent Neural Networks for Short-Term Prediction of Atmospheric Sensing Data, The 4th IEEE International Conference on Smart Data (SmartData 2018), Halifax, Canada, pp. 815-821 (July 2018).