Google Stock prediction using Multivariate LSTM

Not long ago I published a similar article on how to use LSTMs to make Stock predictions using a Vanilla Neural Network. Because I wanted to minimize the complexity of the problem, I used a monovarietal model. Today I will make the use of a multivariate model to train my AI. It will be more complex, but it will begin to be more realistic. The structure I will be using will be almost identical to the one followed in the previous article, with the only difference that this one will be able to incorporate multiple variables (GOOG price and GDP). Real-world models are much more complex, require Multi-variable data and are not limited to a single AI, but rather a collection of AI working together.