Walmart stock price prediction (LSTM vs GRU)

About work

In this project I used two recurrent neural network (RNN) to predict Walmart stock price:

• with long short term memory (LSTM)
• with gated recurrent unit (GRU)

Also, model inference was compared (LSTM vs GRU)

Data

Dataset was collected from investing.com

It consists data from 2013 to 2022 year (Walmart daily stock price values and other additional features)

If you want to use it, you can find my dataset here

LSTM vs GRU

For each of this units, I've trained a simple RNN using MAE loss and Adam optimizer in 50 epochs

After training both of networks inference (prediction + 30-day forecast) was compared

In this case, LSTM works better (MAE = 1.42) than GRU (MAE = 2.49)

Demo

LSTM inference:

GRU inference: