ago
0 like 0 dislike
0 like 0 dislike
So I'm currently working on a deep learning project, and my goal is to forecast power prices one month ahead. I have created my own data set consisting of power price data from Montel, gas-prices, weather data etc, and I want to use these variables in a LSTM-network. Is there anyone who have any experience with creating multivariate LSTM-networks? Do anyone know of any good tutorials on this? Is coding multivariate networks a lot more hassle than univariate? I'm using R with keras/tensorflow.

I will highly appreciate any input, as this is my first time creating a neural network, and my knowledge on the matter right now is rather scarce.

Thank you!
ago
0 like 0 dislike
0 like 0 dislike
I have worked on multivariate time series forecasting using python with tensorflow/keras. I think if this is your first time creating lstm neural network then you should first understand basics of lstm and how it works, after that you can watch any multi variate time series forecasting tutorial or refer to any GitHub repo related to the same.
ago
0 like 0 dislike
0 like 0 dislike
Well, if you’ve never worked with lstm before, I suggest going the univariate route.

Thing is with energy prices, they are highly cyclical and easy to forecast with more naive approaches. You could start with a univariate model and work up from there. If this is for making money and not some assignment, then you’ll probably get more ROI from your time just throwing R’s autoarima at it and moving on.
ago

No related questions found

33.4k questions

135k answers

0 comments

33.7k users

OhhAskMe is a math solving hub where high school and university students ask and answer loads of math questions, discuss the latest in math, and share their knowledge. It’s 100% free!