Are you familiar with the Excel function forecast.ets, for level, trend, and seasonality? I’d like to use it in our forecast. To get buy in, I’d like to prove it out by hand against the result given by the function in Excel. I can get close, but there must be underlying math it uses to choose the parameters.
Hi Austin, sorry, I have never used the Exponential Triple Smoothing (ETS) forecasting function in MS Excel. But send me your model through www.OperationsUniversity.Org and we can try to check the formula & forecast accuracy? If it's a good forecasting tool, maybe we can create a video about it in the future!
Hi Begufi, picking the right Alpha requires a little bit of trial and error based on the forecast volatility at your company and/or for your product. You can pick the Alpha, which is the smoothing constant. The smoothing constant can range from 0 to 1. When the Alpha is close to Zero (0), the smoothing happens more gradually and slowly. When the Alpha is close to One (1), the forecast will be more responsive to changes. Some forecasts won't have much variation (so use an alpha closer to 0), while other forecasts will be more volatile (so use an alpha closer to 1). I hope that helps!
Are you familiar with the Excel function forecast.ets, for level, trend, and seasonality? I’d like to use it in our forecast. To get buy in, I’d like to prove it out by hand against the result given by the function in Excel. I can get close, but there must be underlying math it uses to choose the parameters.
Hi Austin, sorry, I have never used the Exponential Triple Smoothing (ETS) forecasting function in MS Excel. But send me your model through www.OperationsUniversity.Org and we can try to check the formula & forecast accuracy? If it's a good forecasting tool, maybe we can create a video about it in the future!
I will do that!
the only thing i dont get it is alpha factor. How do you decide which alpha factor is the right one?
Hi Begufi, picking the right Alpha requires a little bit of trial and error based on the forecast volatility at your company and/or for your product. You can pick the Alpha, which is the smoothing constant. The smoothing constant can range from 0 to 1. When the Alpha is close to Zero (0), the smoothing happens more gradually and slowly. When the Alpha is close to One (1), the forecast will be more responsive to changes. Some forecasts won't have much variation (so use an alpha closer to 0), while other forecasts will be more volatile (so use an alpha closer to 1). I hope that helps!