Financial Fraud Identification Based on Stacking Ensemble Learning Algorithm: Introducing MD&A Text Information

Table 6

The influence of each text index on the identification results.

Variable	Accuracy	Precision	Recall	F1-score	AUC
F + sentiment polarity	0.8544	0.8298	0.8478	0.8387	0.8567
F + emotional tone	0.8252	0.8200	0.8200	0.8200	0.8251
F + text readability	0.8641	0.8636	0.8261	0.8444	0.8604