Interactive fault localization leveraging simple user feedback

摘要: Many fault localization methods have been proposed in the literature. These take a set of program execution profiles and output list suspicious elements. The elements ranked by their suspiciousness is then presented to developers for manual inspection. Currently, are batch process where developers' inspection efforts rarely utilized ranking. inaccuracy static nature existing prompt us incorporate user feedback improve accuracy methods. In this paper, we propose an interactive framework that leverages simple feedback. Our only needs users label statements examined as faulty or clean, which does not require additional effort than conventional non-interactive After our incorporates such information re-orders rest elements, aiming expose truly earlier. We integrated solution with three well-known methods: Ochiai, Tarantula, Jaccard. evaluation on five Unix programs Siemens test suite shows achieves significant improvements accuracy.