从Github上整理了几个关于时间序列预测的项目,简单记录一下:
@@@tensorflow
Time-series-prediction:Codebase for “Time-series prediction” with RNN, GRU, LSTM and Attention
https://github.com/jsyoon0823/Time-series-prediction
注:Attention class + GRU
@@Keras
Keras Attention Mechanism:Attention mechanism Implementation for Keras.
使用Keras实现 基于注意力机制(Attention)的 LSTM 时间序列预测
https://github.com/philipperemy/keras-attention-mechanism
使用Keras实现 基于注意力机制(Attention)的 LSTM 时间序列预测
@tensorflow
convlstm_Internet_traffic_prediction:
https://github.com/EchoQer/convlstm_Internet_traffic_prediction
注:Encoder+ConvLSTMCell+Decoder
@tensorflow
AR-LSTMs:Predicting Transportation Demand based on AR-LSTMs Model with Multi-Head Attention
https://github.com/ncu-dart/AR-LSTMs
注:encoder_rnn_lstm+decoder_rnn_lstm+multihead_attention
@
LSTM_Attention:
https://github.com/ningshixian/LSTM_Attention
注:各种Attention的model
@@
Stock Prediction Model using Attention Multilayer Recurrent Neural Networks with LSTM Cells:
https://github.com/shoumo95/Stock_Prediction_Model_using_Attention_Multilayer_RNN_LSTM
注:这里的注意力机制使用Tensorflow中的AttentionCellWrapper实现。
attn_cell =tf.contrib.rnn.AttentionCellWrapper(cell=stacked_cell,attn_length=attn_win_size,state_is_tuple=False)
@Pytorch
Emotions-Predictions: Emotions prediction using RNN with LSTM cells and Attention Mechanisms.
https://github.com/ismail-mebsout/Emotions-Predictions
注:EncoderRNN+Attention Decoder