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

Table 2

The performance of the DT model.

Feature extraction method	Matrix size	Testing performance				Cross-validation performance
		Accuracy	Precision	Recall	F1-score	Accuracy	Precision	Recall	F1-score
Unigram	1000	78.91	78.91	78.91	78.48	82.85 ± 0.37	83.32 ± 0.49	82.89 ± 0.35	82.38 ± 0.56
	3000	80.31	80.09	80.31	80.06	87.09 ± 0.75	87.14 ± 0.66	87.15 ± 0.66	86.84 ± 0.66
Bigram	1000	78.34	78.32	78.34	77.96	82.17 ± 0.22	82.72 ± 0.29	82.09 ± 0.28	81.78 ± 0.26
	3000	81.13	80.91	81.13	80.88	85.76 ± 0.5	85.86 ± 0.45	85.86 ± 0.47	85.59 ± 0.58
Trigram	1000	77.92	77.92	77.92	77.53	82.23 ± 0.53	82.84 ± 0.52	82.25 ± 0.48	81.93 ± 0.47
	3000	80.31	80.1	80.31	80.09	86.23 ± 0.87	86.13 ± 0.86	86.09 ± 0.81	85.98 ± 0.87
Four-gram	1000	77.97	77.96	77.97	77.6	81.32 ± 0.59	81.93 ± 0.43	81.37 ± 0.49	81.14 ± 0.53
	3000	80.37	80.15	80.37	80.09	85.73 ± 0.75	85.66 ± 0.74	85.73 ± 0.82	85.45 ± 0.72