I have (3) data sets (each one from independent businesses like mine) for last year, where they enter a seasonal change of business in the next month. I have 1 data set for this year, leading up to this change. How can I use last year's data sets to forecast the expected seasonal change, dependent upon last year's seasonal impact?
Run a multiple regression with the 3 data sets as independent variables and this years data set as the dependent variable and then run a polynomial forecast to keep the seasonal trends of the 3 independent businesses in the forecast.
For a more complex, but accurate method, detect the seasonality within the 3 independent datasets combined and then apply this trend to a linear forecast of the dependent dataset. Several methods exist to detect seasonal trends in time series data. One popular method is X-12-ARIMA which is freely available as a Window executable from the U.S. Census Bureau at http://www.census.gov/srd/www/x12a/x12down_pc.html.
