What is the concept of What-If Analysis?

What-if analysis is a useful feature for model explainability in the AI & Analytics Engine, which allows you to check a model's prediction quickly.

With this feature, you can specify the feature values for an input instance and sense-check the model’s prediction by comparing it with the real target value, if it is known. You can also quickly check the prediction of a model for an unknown instance instead of having to upload a file. As seen in the following screenshot, with the given values for input columns, the instance has a 98% chance of being B (benign).

what-if analysis model prediction

You can also compare the prediction results between different input scenarios. With the example above, decreasing the value of area_worst from 734.6 to 300 reduces the chance of being M (malignant) by around 4% (see below).

What-if analysis input change