An Efficient Sentiment Classification Method with the Help of Neighbors and a Hybrid of RNN Models

Table 4

The effect of neighborhood size.

Neighborhood size	Metrics	Dataset
		BG	BIS	MVSA-Single	Twitter	IMDB
0	Accuracy	98.23	97.53	80.74	95.17	87.01
	Precision	98.41	98.30	80.22	95.16	91.14
	Recall	98.16	96.21	80.74	95.15	83.29
	F1-score	98.10	97.95	80.10	95.11	87.00
2	Accuracy	98.74	98.05	82.28	95.65	87.15
	Precision	98.95	98.78	81.71	95.67	91.32
	Recall	98.67	96.73	82.28	95.66	83.78
	F1-score	98.62	98.46	81.62	95.63	87.12
4	Accuracy	99.23	98.53	82.77	96.15	88.25
	Precision	99.44	99.29	82.20	96.17	92.54
	Recall	99.18	97.24	82.78	96.16	84.34
	F1-score	99.10	98.97	82.10	96.13	88.31
8	Accuracy	99.34	98.65	82.87	96.24	88.36
	Precision	99.55	99.38	82.31	96.28	92.64
	Recall	99.27	97.33	82.87	96.24	84.47
	F1-score	99.22	99.06	82.23	96.22	88.40
16	Accuracy	99.37	98.72	82.94	96.31	88.40
	Precision	99.58	99.40	82.37	96.31	92.68
	Recall	99.31	97.40	82.93	96.32	84.50
	F1-score	99.25	99.10	82.26	96.27	88.44
32	Accuracy	99.42	98.73	82.96	96.33	88.45
	Precision	99.63	99.46	82.39	96.35	92.75
	Recall	99.35	97.41	82.96	96.34	84.12
	F1-score	99.30	99.14	82.30	96.31	88.12