Predicting Coronavirus Pandemic in Real-Time Using Machine Learning and Big Data Streaming System

Table 6

The performance of the SVM model.

Feature extraction method	Dataset	Testing performance				Cross-validation performance
		Accuracy	Precision	Recall	F1-score	Accuracy	Precision	Recall	F1-score
Unigram	1000	79.66	82.38	79.66	78.52	85.63 ± 0.48	87.03 ± 0.51	85.63 ± 0.48	85.26 ± 0.52
	3000	81.17	83.33	81.17	80.35	88.8 ± 0.32	89.52 ± 0.37	88.8 ± 0.32	88.54 ± 0.36
Bigram	1000	79.43	82.28	79.43	78.26	85.07 ± 0.48	86.63 ± 0.53	85.07 ± 0.48	84.65 ± 0.53
	3000	80.79	83.08	80.79	79.95	88.44 ± 0.4	89.19 ± 0.45	88.44 ± 0.4	88.15 ± 0.44
Trigram	1000	79.61	82.25	79.61	78.47	85.04 ± 0.5	86.6 ± 0.52	85.04 ± 0.5	84.62 ± 0.55
	3000	80.78	83.07	80.78	79.93	88.43 ± 0.37	89.19 ± 0.45	88.43 ± 0.37	88.15 ± 0.42
Four-gram	1000	79.62	82.29	79.62	78.49	84.56 ± 0.55	86.27 ± 0.54	84.56 ± 0.55	84.14 ± 0.59
	3000	80.5	82.8	80.5	79.67	88.33 ± 0.37	89.09 ± 0.44	88.33 ± 0.37	88.04 ± 0.42