. 2021 Oct 19;7(10):e08216. doi: 10.1016/j.heliyon.2021.e08216

Table 11.

Comparing F₁-score of deep learning models (TREC-6 dataset) with focusing POS tags (N+V+Adj+Det: NVAD).

Input	Model	Micro-averaged					Macro-averaged
Input	Model	P	R	$F_{1}$ (Mean)	$F_{1}$ (SD)	$F_{1}$ (Max)	P	R	$F_{1}$ (Mean)	$F_{1}$ (SD)	$F_{1}$ (Max)
Original	CNN	0.9600	0.9600	0.9600	0.0042	0.9680	0.9641	0.9652	0.9641	0.0037	0.9710
	BiLSTM	0.9732	0.9732	0.9732	0.0041	0.9800	0.9753	0.9760	0.9755	0.0038	0.9816
	Hybrid	0.9718	0.9718	0.9718	0.0027	0.9780	0.9737	0.9746	0.9741	0.0024	0.9796
+all POS Tags	CNN	0.9594	0.9594	0.9594	0.0031	0.9640	0.9628	0.9641	0.9630	0.0029	0.9674
	BiLSTM	0.9753	0.9753	0.9738	0.0035	0.9780	0.9772	0.9775	0.9758	0.0033	0.9799
	Hybrid	0.9720	0.9720	0.9720	0.0028	0.9760	0.9739	0.9748	0.9742	0.0025	0.9779
+focusing Tags	CNN	0.9644	0.9644	0.9644	0.0034	0.9700	0.9675	0.9688	0.9678	0.0029	0.9723
	BiLSTM	0.9738	0.9738	0.9738	0.0033	0.9780	0.9756	0.9764	0.9759	0.0030	0.9799
	Hybrid	0.9694	0.9694	0.9694	0.0035	0.9740	0.9715	0.9729	0.9720	0.0030	0.9760
+focusing Words	CNN	0.9668	0.9668	0.9668	0.0045	0.9740	0.9697	0.9711	0.9699	0.0041	0.9762
	BiLSTM	0.9724	0.9724	0.9724	0.0040	0.9780	0.9745	0.9753	0.9748	0.0035	0.9798
	Hybrid	0.9750	0.9750	0.9750	0.0056	0.9820	0.9767	0.9779	0.9770	0.0051	0.9834