Table 4.

Summary chart showing the findings from the XAI tools.

Visualization plot	XAI tool	Global/local	Result and key symptoms	Medical significance
Global bar plot	SHAP	Global	“Dry-cough,” “Tiredness,” “Fever,” “Nasal-congestion,” and “Diarrhea”	Overall model behavior; Feature importance ranking;Magnitude of contribution;Verification of domain knowledge
Local bar plot	SHAP	Local	For a selected patient, “Difficulty in Breathing” is found	Explanation of individual predictions; Positive and negative contributions; Feature impact magnitude.
Beeswarm plot	SHAP	Global	“Dry-cough,” “Runny-Nose,” “Tiredness,” “Nasal congestion,” and “Difficulty-in-Breathing”	Interaction effects among features;Outlier detection;Distribution of contributions.
Force plot	SHAP	Local	Positively contributing features for a selected patient are “Runny-Nose” and “Sore Throat”	Granular explanation of an instance-specific breakdown of feature contributions; Net impact calculation indicating the total impact of all features on the prediction.
Waterfall plot	SHAP	Local	For the selected patient, “Difficulty in Breathing” is observed	Visualizing Cumulative Impact of how individual components contribute; Identifying Key Drivers; Forecasting and planning.
Local bar plot	LIME	Local	“Sore-throat,” “Difficulty-in-Breathing,” and “Pains” as the highest positive contribution to the selected patient	Explain individual predictions; Feature impact magnitude (both positive and negative); Model consistency; Verification of domain knowledge.
Violin plot	LIME	Local	Overall feature importance spread per patient	Depicting the distribution of feature importance values generated by LIME; density of the distribution indicates frequent importance values.