Recurrent Neural Network Signal Modeler

A general purpose signal modeler built using the Tensorflow Recurrent Neural Network module.

https://medium.com/towards-artificial-intelligence/signal-modelling-using-recurrent-neural-networks-d832f0c50503

Instructions

1. Define the signal which you would like to model in data_generator.py on line 25
2. Set the directory in which to save the generated data in data_generator.py on line 17
3. Run the data generator (data_generator.py) to generate the signal will be modelled
4. Set the base directory for the recurrent neural network in rnn_signal_modeller.py on line 30
5. Set the hyper-parameters of the network in rnn_signal_modeller.py in line 335-339
6. Run the recurrent neural network (rnn_signal_modeller.py) and wait (execution time depends on the hyper-parameters)
7. Experiment!

Note: By default rnn_signal_modeller will generate intermediate plots across the iterations in base_directory/plots/