LSTM training error is very high and relatively unchanging

Question

As a learning exercise, I\'m trying to use an LSTM model with the Keras framework to predict the stock market based on multiple data points. The size of my input array is roughl