Table 1.

Sentiment polarities in media communications (Google News headlines or descriptions) analyzed using the TextBlob, VADER, and Word2Vec-BiLSTM models.

Period in 2020	TextBlob					VADERa					Word2Vec-BiLSTMb
	Positive	Negative	Neutral	Subjectivity score	Positive		Negative	Neutral	Compound score	Positive		Negative
February 2	77.78	22.22	0.00	0.48	0.07		0.07	0.86	0.02	88.89		11.11
February 9	16.67	83.33	0.00	0.58	0.06		0.08	0.86	–0.01	16.67		83.33
February 16	66.67	0.00	33.33	0.73	0.03		0.15	0.82	–0.35	66.67		33.33
February 23	100.00	0.00	0.00	0.58	0.11		0.15	0.74	–0.11	0.00		100.00
March 1	66.67	26.67	6.67	0.43	0.10		0.08	0.82	0.09	80.00		20.00
March 8	72.73	0.00	27.27	0.64	0.06		0.12	0.82	–0.21	18.18		81.82
March 15	49.18	19.67	31.15	0.40	0.06		0.07	0.87	–0.04	36.07		63.93
March 22	89.74	5.13	5.13	0.37	0.08		0.03	0.89	0.17	69.23		30.77
March 29	53.42	6.85	39.73	0.48	0.07		0.07	0.86	0.02	69.86		30.14
April 5	57.97	17.39	24.64	0.43	0.07		0.05	0.88	0.07	44.93		55.07
April 12	49.09	21.82	29.09	0.34	0.08		0.07	0.85	0.03	70.91		29.09
April 19	75.93	7.41	16.67	0.42	0.06		0.05	0.89	0.10	50.00		50.00
April 26	48.65	32.43	18.92	0.46	0.06		0.10	0.85	–0.14	37.84		62.16
May 3	53.17	17.46	29.37	0.33	0.08		0.07	0.85	0.06	65.08		34.92
Average polarity	62.69	18.60	18.71	0.48	0.07		0.08	0.85	–0.02	51.02		48.98

^aVADER: Valence Aware Dictionary and Sentiment Reasoner.

^bBiLSTM: bidirectional long short-term memory.