Conclusion

Let us summarize what we have learnt so far.

We understood the need of a Recurrent Neural Network which is suited for sequential data and how the long term dependency problem in RNN is solved by introducing Long Short Term Memory nets (LSTM) and another variant GRU.

We were able to predict the stock price of Apple for the past 60 days using an LSTM network and also tried out Bidirectional LSTM to see the improvements in the overall prediction trend.