Twitter Spam dataset (test) benchmark leaderboard · SOTA2 Research

Twitter Spam dataset (test)

94.6Clean AccuracyStd. Training

Evaluation Results

Method	Links
Std. Training2021.02		94.6	17.5
NU-δ-Pearson2021.02		94	36
NU-δ-SHAP2021.02		93.9	35.3
NU-δ-MD2021.02		93.9	36.7
NU-δ-MDtarget2021.02		93.9	38.3
Uniform-δ2021.02		92.6	58.3
NU-δ-MDtarget2021.02		92	67.9
NU-δ-SHAP2021.02		91.9	66.5
NU-δ-MD2021.02		91.9	69.4
NU-δ-Pearson2021.02		91.8	65
Uniform-δ2021.02		91.3	82.8
NU-δ-Pearson2021.02		91.2	87.3
NU-δ-MDtarget2021.02		91.2	86.8
Uniform-δ2021.02		91.1	34.4
NU-δ-MD2021.02		91.1	85.3
Uniform-δ2021.02		91.1	89.6
NU-δ-SHAP2021.02		90.9	86.1
NU-δ-Pearson2021.02		90.6	90.7
NU-δ-SHAP2021.02		90.5	90.5
NU-δ-MD2021.02		90.5	89.8
NU-δ-MDtarget2021.02		90.5	89.1
Uniform-δ2021.02		90.5	87.3
NU-δ-Pearson2021.02		89.9	92
NU-δ-SHAP2021.02		89.8	91.4
NU-δ-MDtarget2021.02		89.8	92.5
NU-δ-MD2021.02		89.7	93.3