Forecast Explanations for Predictive Forecaster

Preview: 2025-02-08

We add a new feature, Forecast Explanation
, to Machine Learning Predictive Forecaster. After your forecast runs with the feature enabled, we provide a chart and explanatory text enabling you to review and understand the factors that contributed to the forecasted data.

Business Benefits

Forecast Explanation helps demystify machine learning predictions. Visual representations make it easier for you to review and understand the factors that influenced each prediction.

Changes

We provide a new option, Forecast Explanation
, in the Forecast
section of new and edited forecasts. When you enable the Forecast Explanation
check box, we provide these new items in the Confidence Metrics
tab of the Forecast History
page:

Contribution Breakdown of Forecast Components
: A visual representation of how seasonality, trend, and residual components (when applicable) contributed to the forecasted values.

Forecast Explanation
: Text describing the chart components.

You can access the Confidence Metrics
when you view the history of a completed forecast.

What Do I Need to Do?

To enable the feature for existing forecasts:

From the main menu, select Modeling
.

Select Predictive Forecaster
.

Hover over the forecasts in the list to access the More Actions
menu.

Click the 3 dots that display and select Edit
.

In the Forecast
section, select Forecast Explanation
.

Select Run
or Save
to run it later.

To review the forecast explanations:

After the forecast runs successfully, select Modeling
from the main menu.

Hover over the forecasts in the list to access the More Actions
menu.

Click the 3 dots that display and select View History
.

Select the Confidence Metrics
tab.

What Happens if I Do Nothing?

The option is available for all new and existing forecasts.

What's Coming Next

We plan to deliver more details in the explanatory text and more charts to help you understand the influences and factors of the predicted data.