LSTM time series forecasting, predicting future values

Hello, I am currently trying to learn how to perform time series forecasting using Matlab. I have been following this MATLAB guide.

However, my goal to to use LSTM to predict future values rather then compare it to known values. This guide take in a data sample of 500 points, is trained and then predicts the points from 450 to 500. My goal is to have it predict points 501 to 550.

Does anyone know how I can alter the code given in the example to achieve this?

Answer 1

[David Willingham]

Hi Stashu,

Take a look at this updated example:

Time Series Forecasting Using Deep Learning

In it there is a section on "closed-loop forecasting" which should provide you with what you're looking for.