BMC Medical Informatics and Decision Making

Table 7 Confusion table of the classification of tweets in the best ML setting (SVM trained on Polarity labeling with strict data appended by lax data)

From: Monitoring stance towards vaccination in twitter messages

		Truth (Strict)
		Irrelevant	Negative	Neutral	Positive	Not clear
Predicted (Best ML)	Irrelevant	172	17	20	60	25
	Negative	74	161	42	230	57
	Neutral	108	37	118	133	55
	Positive	195	103	140	832	84
	Not clear	84	25	25	57	32

The vertical axes give gold standard labels, the horizontal axes give the classifier decisions. Numbers given in bold are accurate classifications

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com