Whether you are a programmer required to provide an accurate forecast or a seasoned forecaster looking to expand your model, Procast can provide an accurate and customizable “best-fit” selection for you.

Procast provides the option to remove or keep outliers, and also provides the option to check for seasonality. Procast takes the kind of error measurement you choose and finds the statistical method that produces the least error to give you results automatically.

Users need only supply the following parameters:

1) Data start point

2) Period to forecast into the future

Procast allows users to specify…

1) the accuracy measure you would like to use. (Default setting is SSE)

2) what seasonality test you would like to perform.

Options available in Procast are:

– Choose from only non-seasonal techniques (use the constant FX_NON_SEASONAL)

– Perform seasonality test and choose from all forecasting techniques except Box Jenkins. (Use the constant FX_OPTIMIZE)

– Perform Seasonality test and choose from all forecasting techniques including Box Jenkins (Use the constant FX_OPTIMIZE_ALL)

– Choose from only seasonal techniques (Use the constant FX_SEASONAL)

Input your own seasonal length of seasonal period ( 1 = Annual, 6 = semiannual, 7 = day of week, 52 = weekly, 12 = monthly, 30 = day of month, 365 = daily)

3) The option to remove outliers is ReplaceOutliers. This effectively removes bad data from the data set.

Procast analyzes your data and chooses the method which will produce the best forecast.

Back to Forecasting Methods