Result visualization methods
Visualization of the model outcomes is a very common approach applied to interpret ML models. Generally, these are model-agnostic, post-hoc analysis methods applied on trained black-box models and provide explainability. In the following section, we will discuss some of the commonly used result visualization methods for explaining ML models.
Using comparison analysis
These are mostly post-hoc analysis methods that are used to add model explainability by visualizing the model's predicted output after the training process. Mostly, these are model-agnostic approaches that can be applied to both intrinsically interpretable models and black-box models. Comparison analysis can be used to produce both global and local explanations. It is mainly used to compare different possibilities of outcomes using various visualization methods.
For example, for classification-based problems, certain methods such as t-SNE and PCA are used to visualize and compare...