LSTM for time series forecasting

In this repository I will implement a LSTM architecture for time series forecasting. We used heartbeat records as dataset.

Requirements

Pandas 0.24.2
Matplotlib 3.0.3
Numpy 1.16.5
Keras 2.2.5

Data

The dataset can be found here.
All data used are under /dataset folder in the main repo.
Here is a visualisation of small sample from the used data.

Results

With only 2 layers of LSTMs (10 units each) we got a quite good results.

The code and the dataset file are availables.
Feel free to contact me for any question.