Anchors provide an intuitive set of rules that emphasis the features that are locally sufficient for the model to make a decision.

These are set of rules in the form of *if-and-then *that are locally **sufficient **in order to ensure a certain prediction with a high degree of confidence. Meaning, that for instances in which the anchors hold, the prediction is almost always the same.

Each anchor consists of the **set of rules, **the **prediction, precision **and **coverage**.

For example:

IF (feature1 >0)

AND (feature2 == x)

THEN Prediction=Blue with Precision 95% and 30% Coverage.

This means that for instances in the local environment of the original sample for which feature1 is positive and feature2 equals to x the prediction will be Blue. This covers 30% of the local samples and for 95% of the instances the model's prediction is correct.

Additional measure for an anchor is its **Complexity**. Complexity is defined to be the number of rules that are used in the anchor. For the example above, the complexity of the anchor is 2.

To learn more about anchors, please refer to __Ribeiro, M. T., Singh, S., & Guestrin, C. (2018, April). Anchors: High-precision model-agnostic explanations. In Thirty-Second AAAI Conference on Artificial Intelligence.__

#explainableai #anchors #scopedrules