Table 10. Comparison of TL and DL results based on accuracy.

DATASETS	SIZE	Transfer Learning					Deep Learning
		ULMFIT	BERT	ROBERTA	ELECTRA	LSTM	CNN
		A	A	A	A	A	A
HOTEL REVIEW	7K	0.928	0.941	0.946	0.918	0.924	0.901
	10K	0.940	0.942	0.946	0.918	0.925	0.898
	15K	0.941	0.948	0.950	0.924	0.924	0.883
	20K	0.952	0.947	0.955	0.931	0.941	0.904
MOVIE REVIEW	7K	0.921	0.917	0.936	0.944	0.801	0.839
	10K	0.921	0.908	0.941	0.936	0.820	0.840
	15K	0.926	0.918	0.935	0.938	0.862	0.857
	20K	0.931	0.911	0.940	0.941	0.885	0.867
SENTIMENT140	7K	0.754	0.804	0.836	0.795	0.688	0.656
	10K	0.770	0.805	0.823	0.801	0.730	0.684
	15K	0.770	0.810	0.834	0.801	0.730	0.691
	20K	0.780	0.814	0.836	0.812	0.757	0.726
CITATION SENTIMENT CORPUS (CSC)	8K	0.910	0.938	0.937	0.932	0.872	0.838
BIOINFORMATICS CITATION CORPUS (BCC)	4K	0.855	0.920	0.894	0.885	0.695	0.693