Table 8 Comparative performance with... | Emerald Publishing

Table 8

Comparative performance with baseline and state-of-the-art models

Model	AUC-ROC	Recall	Type	Characteristics
Logistic regression	0.781	0.724	Statistical	Linear baseline
Decision tree	0.814	0.756	ML	Rule-based
Random forest	0.861	0.812	Ensemble ML	Bagging
LightGBM	0.903	0.845	SOTA	Efficient boosting
CatBoost	0.915	0.858	SOTA	Categorical handling
XGBoost	0.912	0.861	ML	Structured learning
LSTM	0.941	0.903	DL	Temporal modeling
CNN	0.934	0.940	DL	Visual detection
Transformer (TFT)	0.949	0.907	SOTA DL	Long-range modeling
Proposed ensemble	0.963	0.921	Hybrid AI	Multimodal fusion + calibration

Model	AUC-ROC	Recall	Type	Characteristics
Logistic regression	0.781	0.724	Statistical	Linear baseline
Decision tree	0.814	0.756	ML	Rule-based
Random forest	0.861	0.812	Ensemble ML	Bagging
LightGBM	0.903	0.845	SOTA	Efficient boosting
CatBoost	0.915	0.858	SOTA	Categorical handling
XGBoost	0.912	0.861	ML	Structured learning
LSTM	0.941	0.903	DL	Temporal modeling
CNN	0.934	0.940	DL	Visual detection
Transformer (TFT)	0.949	0.907	SOTA DL	Long-range modeling
Proposed ensemble	0.963	0.921	Hybrid AI	Multimodal fusion + calibration

Source(s): Authors’ own work