Table 2.

10-fold cross validation micro-averaged results using bag of 1–4 gram features as baselines (categories abuse, anger, blame and pride are not included to avoid False Positive).

	P	R	F
SVM	0.7767	0.2772	0.4085
SVM (tune threshold of each category)	0.4779	0.5325	0.5038
Naïve Bayes	0.5481	0.4088	0.4683
Boosting*	0.6497	0.3493	0.4543

Notes:

^*A multiclass boosting algorithm²⁶ is used. Weak classifiers: decision trees (depth = 2); number of iterations: 500. The default parameters are used for baseline experiment.