spam-detection's Introduction

spam-dectection

Implement of paper Learning to Represent Review with Tensor Decomposition for Spam Detection,emnlp,2016

results

RE(50:50):0.7792207792207793

            precision    recall    f1-score     support
      0      0.75      0.83      0.79         77    
      1      0.81      0.73      0.77         77   
  avg/total    0.78      0.78      0.78         154 


RE(ND):0.7123050259965338

              precision    recall  f1-score   support
        0       0.29      0.83      0.44        77
        1       0.96      0.69      0.81       500
  avg/total     0.87      0.71      0.76       577



RE+PE(50:50):0.811688311688

               precision    recall  f1-score   support
          0       0.78      0.87      0.82        77
          1       0.85      0.75      0.80        77
avg / total       0.82      0.81      0.81       154



RE+PE(ND):0.738301559792

         precision    recall  f1-score   support
        0       0.32      0.83      0.46        77
        1       0.97      0.72      0.83       500
avg/total       0.88      0.74      0.78       577



RE+PE+BiGram(50:50):0.746753246753

              precision    recall  f1-score   support
          0       0.72      0.82      0.76        77
          1       0.79      0.68      0.73        77
  avg/total       0.75      0.75      0.75       154



RE+PE+BiGram(ND):0.769497400347

              precision    recall  f1-score   support
         0       0.35      0.83      0.49        77
         1       0.97      0.76      0.85       500
  avg/total      0.88      0.77      0.80       577

Recommend Projects