LSTM is a time recurrent neural network (RNN), mainly to solve the problems of gradient disappearance and gradient explosion in the process of long sequence training. In short, compared with ordinary RNN, LSTM can perform better in longer sequences.
Input variable: time series data
Output variable: predicted data
Contact SupportNorthwest Institute of Eco-Environment and Resources, CAS 0931-4967287 email@example.com
LinksNational Tibetan Plateau Data Center