Table 1.

Baseline performance of LR9 and ADTree7 on 420 ratings of 140 YouTube videos (average performance and standard deviation). With listwise deletion, a rating is dropped if it contains at least one NULL value in the model's features (this is the case for 135 ratings for LR9). We are unable to rate a video if at least one model feature is missing in all 3 ratings of this video (this is the case for 5 videos for LR9). 

Model	Sensitivity	Specificity	UAR	AUC-ROC	AUC PR	Ratings dropped
LR9	0.8939 (0.0152)	0.7536 (0.0102)	0.8238 (0.0118)	0.9109 (0.0021)	0.9195 (0.0018)	135 ratings and 5 videos entirely
ADTree7	0.8172 (0.0397)	0.8721 (0.0709)	0.8447 (0.0261)	0.9083 (0.0117)	0.8706 (0.0644)	205 ratings and 21 videos entirely