Skip to main content

Table 13 Classification of author intent in the HF2200 dataset: H1 confusion matrix

From: Automatically identifying the function and intent of posts in underground forums

Prediction Reference
Neutral Negative Positive Arbitrate Aggression Private-message Gratitude Vouch Total
Neutral 1504 324 807 45 22 107 13 26 2848
Negative 13 16 9 1 0 0 0 1 40
Positive 93 33 130 2 3 6 12 7 286
Arbitrate 9 12 9 25 0 1 0 1 57
Aggression 20 29 15 2 23 0 0 0 89
Private-message 114 31 89 5 7 20 12 12 290
Gratitude 68 9 33 2 1 11 97 3 224
Vouch 34 8 52 0 0 6 18 52 170
Total 1855 462 1144 82 56 151 152 102 4004