Skip to Main Content
Table 15 

Results for our three models on the Chinese data sets (one consisting of raw documents and the other of generated documents), using continuous classifier outputs. Best result per column in bold, best result per row grouping is underlined.

 FeatureAccuracy (%)
rawgenerated
Baselines Random Baseline 9.1 9.1 
Majority Class Baseline 16.1 12.9 
Single Vector 44.7 70.6 
Current Best Result n/a 70.6 
  
Oracles Oracle 67.7 92.2 
Accuracy@2 57.1 76.9 
Accuracy@3 65.3 84.7 
  
Ensembles Plurality Voting 43.5 68.5 
Borda Count 43.2 66.4 
Mean Probability 45.4 71.1 
Median Probability 43.4 66.1 
  
Meta-classifier Linear SVM 49.5 75.4 
Logistic Regression 49.5 74.6 
Ridge Regression 48.3 71.4 
LDA 50.8 75.9 
  
Meta-classifier Linear SVM 49.6 75.5 
Bagging Logistic Regression 49.5 75.1 
Ridge Regression 48.4 71.4 
LDA 51.2 76.5 
 FeatureAccuracy (%)
rawgenerated
Baselines Random Baseline 9.1 9.1 
Majority Class Baseline 16.1 12.9 
Single Vector 44.7 70.6 
Current Best Result n/a 70.6 
  
Oracles Oracle 67.7 92.2 
Accuracy@2 57.1 76.9 
Accuracy@3 65.3 84.7 
  
Ensembles Plurality Voting 43.5 68.5 
Borda Count 43.2 66.4 
Mean Probability 45.4 71.1 
Median Probability 43.4 66.1 
  
Meta-classifier Linear SVM 49.5 75.4 
Logistic Regression 49.5 74.6 
Ridge Regression 48.3 71.4 
LDA 50.8 75.9 
  
Meta-classifier Linear SVM 49.6 75.5 
Bagging Logistic Regression 49.5 75.1 
Ridge Regression 48.4 71.4 
LDA 51.2 76.5 
Close Modal

or Create an Account

Close Modal
Close Modal