A Fault Localization Method Based on Metrics Combination

Spectrum-Based Fault Localization (SBFL) is one of the most effective fault localization techniques, and its performance closely depends on program spectra ranking formula. Despite numerous proposed approaches for localization, there are still great demands techniques that can help guide developers to locations faults. Therefore, this paper defines four metrics from spectrum, which become essential components formulas mitigate spectrum-based problems. These further combined propose a new heuristic, Metrics Combination (MECO), does not require any prior information structure or semantics locate faults effectively. The evaluation experiments conducted Defects4J SIR datasets, MECO compared with 18 maximal formulas. experimental result shows more efficient in terms Precision, Accuracy, Wasted Efforts than An empirical also indicates two defined metrics, Assumption Proportion Assumption, when existing formulas, improve effectiveness, especially precision ER5a-c (77.77%), GP02 (41%), GP19 (27.22%), respectively.