LSTM - Window Method

Hi all

I would like to build a LSTM for forecasting using a dataset where I have values for each date/hour like below:

2021-01-04 00:00 AM;0.00042
2021-01-04 01:00 AM;0.00375
2021-01-04 03:00 AM;0.00021
2021-01-04 05:00 AM;0.00164
2021-01-04 06:00 AM;0.06367
2021-01-04 07:00 AM;0.05686
2021-01-04 08:00 AM;0.816
2021-01-04 09:00 AM;1.84477
2021-01-04 10:00 AM;2.14783
2021-01-04 11:00 AM;1.87455
2021-01-04 12:00 AM;1.3206
2021-01-04 01:00 PM;1.41993
2021-01-04 02:00 PM;1.76163
....... until 2021-30-04 11:00 PM

My data started in jan 2021 until now - month by month.
I want to predict the next month. Initially the full month but if predict only first 15 days is better, no problem… after i can predict the next 15 days.

Somebody has an example with code (link… blog… article) where I can learn the “windows method” using LSTM in Tensoforflow?

Perhaps this article Perkiraan deret waktu  |  TensorFlow Core can help you.

1 Like

Thanks Yasir for your suggestion!