Logic of Implementaion

#1

Hi Vasu Bhaiya,
I wanted to confirm my approach.
I was thinking to implement a LSTM layer with 19 cells.
Then I would set the return sequences as true.
This way I plan to make the model.

The doubts are

  1. Do I need to use a functional Api for getting multiple outputs. Where each cell gets a input of 1 high price and it predicts the price for next day.

  2. the units we mention in the LSTM layer are basically dimensions of output space …So this is the dimension of the state vector right ?

As y wil have a different output dimension as compared to a

#2

Hey @chiragwxN, this is one way of dealing, the other way around is, pass the previous say, 20/30/50 values to the model keeping return sequences = False and predicting one value as output. means everytime you predict you have to pass previous n values. Also, for test data values are continuous, means you have to append all test data in traning data at last, pass previous n values for each test data and predict the output values.

  1. Do I need to use a functional Api for getting multiple outputs. Where each cell gets a input of 1 high price and it predicts the price for next day.

Yes you have to use functional model.

  1. the units we mention in the LSTM layer are basically dimensions of output space …So this is the dimension of the state vector right ?

Yes the units we specify is the dimension of output space. Number of lstm cells is equal to the max_length or number of previous values you pass to the model.

Hope this cleared your doubt.
Plz mark the doubt as resolved in my doubts section.

#3

I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.

On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.