Research and Practice in Technology Enhanced Learning

Table 5 Doubt sentic pattern distribution on both training and testing data (‘n’ is data labelled as having no doubt and ‘y’ as having doubt)

From: Automated doubt identification from informal reflections through hybrid sentic patterns and machine learning approach

Pattern	‘n’ labelled data	‘y’ labelled data	Difference
Questioning	0.2%	15.1%	75.5
Confirming	3%	11.6%	3.87
Affirming	5%	27.1%	5.42
Clarifying	4%	10.8%	2.7
Dissenting	5%	27.5%	5.5

Back to article page