22
Results:Classifiers
§
§
95.89
92.65
95.88
F (%)
95.81
91.39
97.50
P (%)
95.98
93.91
94.25
R (%)
SVM   (RBF)
SVM (Linear)
Decision Tree
Using all three feature groups:
Best result: SVM with Radial Basis Function
(0,5)
As we can see, the decision tree and SVM with radial basis function give comparable performance in terms of the f measure, and both are much better than SVM with linear kernel. (These are measures when the highest f measure is achieved within each setting.) Also, In this particular setting,  SVM with radial basis function gives similar precision and recall, while decision tree achieves a better precision but a worse recall.