Table 8 Evaluation of the classification algorithms: overall accuracies based on 10-fold cross-validation on the training sets of the Android and Swing
From: Redocumenting APIs with crowd knowledge: a coverage analysis based on question types
Tool | Classification | Android | Swing | Android + |
|---|---|---|---|---|
algorithm | (%) | (%) | Swing (%) | |
Weka | IBk | 73.35 3B | 76.33 3B | 74.79 1A |
J48 | 73.02 1B | 78.56 1A | 73.31 3B | |
NaiveBayes | 68.21 2B | 74.66 2A | 71.67 3B | |
BayesNet | 73.35 23AB | 75.23 1AB | 74.65 1AB | |
DecisionTable | 70.83 3AB | 77.73 3AB | 73.02 1B | |
MultilayerPercep. | 73.97 3B | 76.06 1B | 76.44 1B | |
SMO | 72.7 3B | 75.24 2A | 75.98 2B | |
RandomForest | 73.66 1A | 78.02 1A | 75.54 3A | |
SimpleLogistic | 78.45 2B | 78.28 2B | 77.46 2B | |
Logistic | 77.5 3B | 78.83 2B | 78.19 2B | |
Mallet | C45 | 57.14 | 53.47 | 54 |
NaiveBayes | 67.24 | 64.94 | 63.24 | |
MaxEnt | 71.76 | 69.37 | 76.58 | |
MaxEntL1 | 77.5 | 69.95 | 77.01 |