Hello,
I need (link, paper, or book chapter) explains in details (step by step)
by numerical example "How to build a neural network model for uni-variate time series?".
For example:
time series includes
20,30,50,80,10,90,...... and so on.
How to determine the neural network model parameters that need for forecasting the incoming values.
Thanks in advance.
Hi,
Did you try to have a look at the manual of Neural Network Toolbox in MATLAB.
Marwa,
As Ayman mentioned, MATLAB has a pretty helpful Neural Network toolbox - If you're dealing with a timeseries, you're probably going to end up using the NARX model (see link below). Also, unlike regression models, NARX models are not effected due to the higher dimension created by the variables. Thus you would want to ask experts in the field for correlated variables to the response - or identify some variables yourself and perform correlation analysis to pick the correlated variables.
Hope this helps:
Armin
Link: https://www.mathworks.com/help/nnet/ug/design-time-series-narx-feedback-neural-networks.html
This is covered with R-software in our new forecasting textbook, Ord, Fildes and Kourentzes,
Principles of Business Forecasting, Wessex, 2017
It should give you everything you need.
- Badges
- Science topic
- Similar topics
- Control Systems Engineering
- Filtering