ClearBrain automatically scans all your user attributes and actions, and determines their relative importance to performing a goal.
But sometimes you want to leverage your own domain knowledge about which attributes are most important, and evaluate how specific attributes when included or excluded will affect a prediction.
ClearBrain offers an "Advanced Settings" feature which you can use to accordingly customize which user attributes will be used as inputs to your prediction goal.
Step 1: Select Advanced Settings
The first step is to select the Advanced Settings Tab in the Prediction Wizard.
When creating a new Prediction, proceed to the final step, where you will see an "Advanced Settings" option in the bottom right corner.
In selecting the "Advanced Settings" option, you'll be directed to a new screen where you can specify the attributes you want to include or exclude from your prediction.
Step 2: Specify attributes for your Prediction
Within the "Advanced Settings" Menu, you can pick any specific user attribute to focus on in your analysis.
In the center of the page, you'll see a series of filter menu options where you can select any user trait or action. Choose an attribute of interest, and then click "Exclude" or "Include" to select that attribute.
You can search and select for as many attributes as you're interested in analyzing.
Step 3: Choose to Exclude or Include attributes in your Prediciton
In choosing the attributes for your prediction, you can decide between including or excluding the selected attributes.
The default option is to "Exclude" selected attributes. This will in turn use all your user attributes to evaluate a prediction, and explicitly remove those attributes you have selected.
If you toggle the option to "Include Only", the prediction will instead only include the selected user attributes to evaluate the prediction, and exclude all other attributes.
Step 4: Run your prediction
Once you've selected the attributes you would like included or excluded in your prediction, you will be redirected back to your Prediction Details page. Simply confirm the details and run the prediction as you normally would.
By selecting only a subset of attributes to train your prediction, it is possible your prediction's accuracy may decrease. However, selecting a subset of attributes can help you gain more insight into which attributes are more important to user paths in your prediction by reducing noise from other attributes.