Table 5

Conditions of reconstruction agreement. Emotion agreement is shown as F1 and Acc. between pairs of labels from a generator and a validator (G–V) or two validators (V–V). The appraisal agreement is an average root mean square error. #Pairs denotes the number of G–V (1st) or V–V (2nd) pairs for each condition. Boxes indicate measures computed on the same textual instances, which can therefore be directly compared. * indicates all pairs that are significantly different from each other inside a box; calculated with 1,000× bootstrap resampling, confidence level .95.

Conditions of reconstruction agreement. Emotion agreement is shown as F1 and Acc. between pairs of labels from a generator and a validator (G–V) or two validators (V–V). The appraisal agreement is an average root mean square error. #Pairs denotes the number of G–V (1st) or V–V (2nd) pairs for each condition. Boxes indicate measures computed on the same textual instances, which can therefore be directly compared. * indicates all pairs that are significantly different from each other inside a box; calculated with 1,000× bootstrap resampling, confidence level .95.
Conditions of reconstruction agreement. Emotion agreement is shown as F1 and Acc. between pairs of labels from a generator and a validator (G–V) or two validators (V–V). The appraisal agreement is an average root mean square error. #Pairs denotes the number of G–V (1st) or V–V (2nd) pairs for each condition. Boxes indicate measures computed on the same textual instances, which can therefore be directly compared. * indicates all pairs that are significantly different from each other inside a box; calculated with 1,000× bootstrap resampling, confidence level .95.
Close Modal

or Create an Account

Close Modal
Close Modal