I am currently working on a load forecasting project to forecast electricity load demand for one ahead.
I have trained the network by using net=train(net p,t) and i have simulated it using sim(net ,p) comand.
How can I predict 12 months ahead?
I need a sample code.
Hello.
Try the comand nntool, on the workspace.
You don't need to write a code.
You can find some samples on youtube.
Regards.
There are two ways it is possible to do. First, using the estimated parameters for your ANN, fit the model using the lags simply by multiplying them, as in a linear regressión. Second (my favourite, but time-consuming) is to forecast one period ahead, and reestimate the neural network using your first parameters as inicial values and including your forecast, and do it recursively untile the h-step you want to forecast. I implemented it using the "neuralnet" package of R.
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators