4.3 Reference Score Table

The below score table is a reference score table that is derived from the study that was executed. This is a generic/standard score we would like to hold for the Algorithms and the Methods for reference. It still needs examination to finalize and standardize the score to use it as a reference score table for any reader or user who would like to score their model and make their preferences.

4.3.1 Reference Score Table - Interpretable Models/Model-Specific

Algorithm	Linearity	Monotonicity	Interaction	Parametric	Transparent	Algorithmic complexity
Logistic Regression	1	1	1	1	1	1
Generalized Linear	1	1	1	1	1	0.9
Gradient Boosting	0.5	0.7	0.7	1	1	0.8
XGBoost	0.1	0.4	0.7	0.9	0	0.7
Distributed Random Forest	0.1	0.4	0.7	1	0	0.9
Deep Learning	0	0	0	0.1	0	0.2
K-means	0.1	0.8	0.1	0.7	1	0.5

4.3.2 Reference Score Table - Model-Agnostic

Explanation Method	Feature summary	Model internals	Datapoint	A surrogate intrinsically interpretable model	Expressive power	Portability	Algorithmic complexity	Detailed	Correctness	Consistency	Stability	Certainty	Importance	Novelty	Representativeness
Variable Importance / Standardized Coefficient	1	1	0.5	/	1	1	1	1	1	0.3	1	1	1	0	1
PDP	1	0.1	0.9	/	1	1	0.9	0.5	1	0.8	1	0.8	0.8	1	0.1
ICE	1	0.2	1	/	1	0.5	0.5	1	1	0.8	1	0.9	1	0.9	0.1
ALE	1	0	0.9	/	1	1	1	0.6	1	0.9	1	0	1	1	0.1

NOTE: The above can be used for references. Though having said that, this table is yet to be standardized, and socred better with mathematical inferences.