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

Table 4

The performance of the LR model.

Feature extraction method	Matrix size	Testing performance				Cross-validation performance
		Accuracy	Precision	Recall	F1-score	Accuracy	Precision	Recall	F1-score
Unigram	1000	80.82	81.22	80.82	80.38	84.54 ± 0.4	85.53 ± 0.48	84.54 ± 0.4	84.16 ± 0.43
	3000	82.94	83.01	82.94	82.61	89.22 ± 0.4	89.36 ± 0.42	89.22 ± 0.4	89.08 ± 0.41
Bigram	1000	80.8	81.33	80.8	80.33	83.98 ± 0.31	85.11 ± 0.35	83.98 ± 0.31	83.56 ± 0.32
	3000	82.32	82.84	82.32	81.84	88.52 ± 0.38	88.86 ± 0.43	88.52 ± 0.38	88.31 ± 0.39
Trigram	1000	80.56	81.08	80.56	80.09	83.92 ± 0.32	85.04 ± 0.39	83.92 ± 0.32	83.5 ± 0.33
	3000	82.32	82.84	82.32	81.84	88.52 ± 0.38	88.86 ± 0.43	88.52 ± 0.38	88.31 ± 0.39
Four-gram	1000	80.32	80.87	80.32	79.85	83.47 ± 0.31	84.69 ± 0.33	83.47 ± 0.31	83.05 ± 0.33
	3000	82.32	82.35	82.32	82.01	88.36 ± 0.43	88.55 ± 0.46	88.36 ± 0.43	88.18 ± 0.44