Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It only takes a minute to sign up.
Sign up to join this community
Disclaimer: New to this field
I am researching ways to forecast a given time-series.
So, I understand what Univariate and Multivariate forecasting is. While going through different results, I came across multistep time-series forecasting(link) and ways to do that.
My assumption: Whenever you forecast multiple times in a series, its called multi-step. Is this the definition or something much more detailed?
Your assumption is correct, regarding your example. The author forecasts one(t), two(t+1) and three month (t+2) in advance, so he called it multi-step. It is to be expected that the near periods are easier to predict.
There are different possiblilities to do that:
First you could just use your training data to forecast all three time steps, like the author does.
Second, you could forecast the the value for t and then use that new "data" for your prediction in t+1 etc...This might improve your prognosis.
