Knowledge extraction methods
Whenever we talk about explainability in any context, it is all about gaining knowledge of the problem so as to gain some clarity about the expected outcome. Similarly, if we already know the outcome, explainability is all about tracing back to the root cause. Knowledge extraction methods in ML are used to extract key insights from the input data or utilize the model outcome to trace back and map to certain information known to the end users for both structured data and unstructured data. Although there are multiple approaches to extracting knowledge, in practice, the data-centric process of EDA is one of the most common and popular methods for explaining any black-box model. Let's discuss more on how to use the EDA process in the context of XAI.
EDA
I would always argue that EDA is the most important process for any ML workflow. EDA allows us to explore the data and draw key insights; using this, we can form certain hypotheses from the data...