TS Forecast Factory Tool

The TS Forecast Factory tool provides forecasts from groups of either ARIMA or ETS models for a user-specified number of future periods. In addition, upper and lower confidence interval bounds are provided for two different user-specified percentage confidence levels. For each confidence level, the expected probability that the true value will fall within the provided bounds corresponds to the confidence level percentage.

This tool uses the R tool. Go to Options > Download Predictive Tools and sign in to the Alteryx Downloads and Licenses portal to install R and the packages used by the R Tool.

Connect inputs

The TS Forecast Factory tool has two anchors for incoming data.

• M anchor: An Alteryx stream of ARIMA or ETS model objects generated by the TS Model Factory tool. Generally, use the O output from the TS Model Factory tool as the M input.

  Because the TS Model Factory tool produces some models using covariates and others without covariates, it is recommended that you filter these models to use one TS Forecast Factory tool on the models with covariates and one on the models without them.

• C anchor: Optional. The future values for covariate fields. Use this input only if the original models used covariates. The number of rows in this input must equal the product of the number of groups with covariates and the number of periods being forecast.

Configure the tool

The field name for the point forecast: Enter a name for the column to contain the point forecasts for the groups. Point forecasts are the reported value of the forecast, in contrast to the confidence intervals.

The percentage value of the larger confidence interval: Set a value between 1 to 99. Common values are 95 or 99 percent. For example, a 95 percent confidence interval would mean that there is a 95% probability that the true value is within the interval.

The percentage value of the smaller confidence interval: Set a value between 1 to 99. Common values are 80 or 90 percent.

The number of future periods to forecast (assuming no covariates in model): Specify the number of periods to forecast.

View the output

Connect a Browse tool to each output anchor to view results.

• N anchor: A data stream containing forecasts and confidence bounds for models that do not use covariates.
• C anchor: A data stream containing forecasts and confidence bounds for models that do use covariates.