A Deep-Learning Framework for Analysing Students’ Review in Higher Education

Table 8

Comparison of proposed approach in this work with state-of-the-art performance for sentiment polarity prediction.

Approaches	Source	No. polarity categories	Precision	Recall	F1-score	Accuracy
SVM	[28]	3	0.91	0.88	0.91	0.97
NB	[28]	3	0.70	0.70	0.71	0.89
CNB	[28]	3	0.78	0.87	0.89	0.93
ME	[28]	3	0.92	0.82	0.92	0.87
LSTM + domain embedding	[19]	3	0.88	0.85	0.86	—
WS-LSTM + glove	[21]	3	—	—	0.93	—
WS-CNN + glove	[21]	3	—	—	0.91	—
Baseline model	This study	2	0.94	0.94	0.94	0.94
Proposed model	This study	2	0.96	0.96	0.96	0.96