[Retracted] XAI-Based Reinforcement Learning Approach for Text Summarization of Social IoT-Based Content

Table 5

Evaluation results of generated summarization on CNN/daily mail dataset %.

How the abstract is generated	Model method	ROUGE-1	ROUGE-2	ROUGE-L	ROUGE-AVG	METEOR
Extraction	MMS_Text	47.57	25.72	44.42	39.24	26.97
	SummaRuNNer	48.6	25.2	44.3	39.37	26.75
	Refresh	49.27	27.2	45.6	40.69	27.38
	HSSAS	51.3	26.8	46.6	41.57	28.27
Generative	Pointer-generator + coverage	48.53	26.28	45.38	40.06	27.7
	Bottom-up	50.22	27.68	47.34	41.75	28.38
	DCA	50.69	28.47	46.92	42.03	28.55
Supervised	BERTSUMEXTABS	51.13	28.6	48.18	42.64	28.91
	PEGASUSBASE	50.79	27.81	47.93	42.18	28.63
	XAI-RL4 (Table 2 combination 4)	51.64	29.13	49.33	43.37	29.7
	XAI-RL5 (Table 2 combination 5)	52.01	29.28	49.71	43.67	29.48
	XAI-RL6 (Table 2 combination 6)	53.29	30.55	40.13	44.66	20.51