Table 3.

TextNetTopics performance results for the WOS dataset, utilizing various topic modeling methods in the T component (150 features).

Topic Model	Accuracy	Recall	Precision	F1-score	AUC	Specificity
LSI	97.11	98.00	96.26	97.12	99.49	96.23
CTM	97.06	97.37	96.75	97.04	99.35	96.73
PLSA	96.76	97.51	96.04	96.76	99.02	96.02
CombinedTM	96.60	97.05	96.15	96.60	99.34	96.16
LDA	96.38	96.98	95.78	96.37	99.24	95.78
NMF	96.34	97.26	95.47	96.35	99.19	95.43
LDA2VEC	95.92	96.67	95.21	95.93	99.04	95.19
TOP2VEC	95.80	96.35	95.26	95.80	99.01	95.26
ETM	95.45	95.86	95.05	95.44	98.88	95.05
BERTopic	95.44	95.96	94.91	95.43	98.84	94.91
FLSA	94.41	94.67	94.15	94.38	98.65	94.12

The highest values for each performance metric are highlighted in bold.