Skip to main content

Table 4 Logistic Regression Results: One-Versus-All Classification. ROC-AUC scores for one-versus-all logistic regression experiments for all categories. Feature sets consist of the number of revisions (nrev), control variables (CV), network metrics (NM), label frequencies (LF), and label transition probabilities (LT). Utilizing all feature sets performs best for all experiments

From: Relating Wikipedia article quality to edit behavior and link structure

Features ROC-AUC
  LQ vs All HQ vs All CI vs All EW vs All
nrev 0.675 0.605 0.778 0.950
CV 0.720 0.702 0.789 0.970
NM 0.745 0.722 0.809 0.915
LF 0.617 0.729 0.789 0.885
LT 0.691 0.762 0.793 0.898
CV + NM 0.775 0.767 0.821 0.970
CV + LF 0.737 0.754 0.814 0.980
CV + LT 0.782 0.792 0.811 0.982
NM + LF 0.759 0.793 0.858 0.951
NM + LT 0.788 0.806 0.861 0.956
LF + LT 0.709 0.809 0.852 0.948
NM + LF + LT 0.796 0.843 0.887 0.970
CV + NM + LF 0.788 0.809 0.861 0.980
CV + NM + LT 0.814 0.825 0.866 0.981
CV + NM + LF + LT 0.823 0.853 0.888 0.986