-------- BASELINE -------- 70.02667% of 1875 tokens lack PPs -------- SVM OUTPUT -------- Best SVM Accuracy: 78.93333 Average SVM Accuracy: 78.60267 Model Predictions for the Best Model: true pred 0 1 0 1253 234 1 60 328 Observed Accuracy for the Best Model: 84.32 Cohen's Kappa (Human-to-SVM): 0.5901907 -------- SVM (NOVEL DATA) -------- Novel Data Predictions for the Best Model: true pred 0 1 0 256 66 1 12 41 Novel Data Accuracy for the Best Model: 79.2 Novel Data Cohen's Kappa (Human-to-SVM): 0.3988656 -------- RANDOMFOREST OUTPUT -------- RandomForest Accuracy: 82.82667 RandomForest Feature Importance: MeanDecreaseGini areadiff 372.4301 absareadiff 413.9146 -------- NAIVE BAYES OUTPUT -------- Naive Bayes Accuracy: 68.10667 Model Predictions for NaiveBayes: nbpred 0 1 0 999 284 1 314 278 Observed Accuracy for NaiveBayes: 68.10667 Cohen's Kappa (Human-to-NaiveBayes): 0.2516728 -------- ENSEMBLE INFO -------- Best Ensemble Classifier: RandomForest Best Ensemble Accuracy: 82.82667 Best Improvement on Baseline: 12.8 Mean Ensemble Accuracy: 76.62222