Fig 1.
Followers’ domains discovery using Twitter LDA.
Table 1.
The domains identified from followers’ tweets using Twitter LDA.
Fig 2.
The bootstrapping algorithm.
Table 2.
The configuration of bootstrapping using a single SVM model.
Fig 3.
A general architecture of bootstrapping using a single SVM model.
Fig 4.
A general architecture of the ensemble system using multiple SVM models.
Table 3.
The configuration of various multiple SVM ensembles.
Fig 5.
The majority vote algorithm.
Fig 6.
The bagging algorithm.
Fig 7.
The stacking algorithm.
Table 4.
Topic groups identified via the seed words-fuzzy match approach and some of their topical words.
Fig 8.
F measures of 10 SVM models generated from random samples.
The x-axis represents individual SVM models.
Fig 9.
AUC of 10 SVM models generated from random samples.
The x-axis represents individual SVM models.
Table 5.
Results of 10 fold cross-validation of various SVM ensembles.
Table 6.
Results of various SVM ensembles on the testing dataset.
Fig 10.
ROC curves of various SVM ensembles on the testing dataset.