Detailed Performance Evaluation
Here you can find the detailed results of the performance evaluation of approx. The results include measurements obtained using query patterns of different lengths. Each query pattern was taken from a code fragment of the subject systems that was identified to be defective. Moreover, each run has been repeated using different thresholds to tolerate differences for each pattern. That is, a search has been performed for each query pattern three times tolerating 0%, 10%, and 20% differences.
You will find the results of our performance evaluation in the following order:
-
DrJava
- Pattern length: 248 tokens
- Pattern length: 133 tokens
- Pattern length: 121 tokens
- Pattern length: 116 tokens
- Pattern length: 82 tokens
- Pattern length: 67 tokens
- Pattern length: 65 tokens
- Pattern length: 48 tokens
- Pattern length: 45 tokens
- Pattern length: 44 tokens
-
jEdit
- Pattern length: 151 tokens
- Pattern length: 122 tokens
- Pattern length: 113 tokens
- Pattern length: 98 tokens
- Pattern length: 84 tokens
- Pattern length: 80 tokens
- Pattern length: 79 tokens
- Pattern length: 57 tokens
- Pattern length: 54 tokens
- Pattern length: 45 tokens
DrJava
Pattern length: 248 tokens
Pattern length: 133 tokens
Pattern length: 121 tokens
Pattern length: 116 tokens
Pattern length: 82 tokens
Pattern length: 67 tokens
Pattern length: 65 tokens
Pattern length: 48 tokens
Pattern length: 45 tokens
Pattern length: 44 tokens
jEdit
Pattern length: 151 tokens
Pattern length: 122 tokens
Pattern length: 113 tokens
Pattern length: 98 tokens
Pattern length: 84 tokens
Pattern length: 80 tokens
Pattern length: 79 tokens
Pattern length: 57 tokens
Pattern length: 54 tokens
Pattern length: 45 tokens
