Search code examples
time-seriesforecastingintermittent

Is it possible to pass multiple columns at once to Croston method?

I want to implement the Croston method for intermittent demand. I have a data frame that has 10 features and all those have many zeros in them. I want to pass the entire data frame to the Croston model but the model accepts a one-dimensional array. I'm not interested in looping through the model. Is there any way or any other methods that can give forecasts for intermittent demand?.

Thanks in Advance!!

Solution

  • Any other methods that can give forecasts for intermittent demand?

    In order to have the forecasts for the intermittent demand, Croston is one of the best methods in the industry but there are few drawbacks to it and those are addressed by the variants of the Croston model like SBA, TSB etc. For me, Croston TSB did perform much better because it can able to decay towards zeros when there is no demand for a long time.

    Next,

    Passing an entire data frame to the Croston model in one go without looping

    NumPy vectorization is much better in this case. vectorization is way better than loops in terms of speed. You can refer on vectorization.