We propose an approach in which the developer interacts with the SBFL algorithm by giving feedback on the elements of the prioritized list. We exploit contextual knowledge of the user about the next item in the ranked list, with which larger code entities can be repositioned in their suspiciousness.
