Improved Similarity Computation Method of Execution Trace
Abstract
Fault localization is a crucial point for the quality of software test. After analyzing the advantages and disadvantages of fault localization method based on trace, we proposed a new computation method to confirm the correlation pattern of statement blocks through relevance of minor defects in the program. We also introduced this method into computation of vector similarity and established a new computation method for execution trace similarity. With realization of details as the basis, the effectiveness of this new method has been verified in combination with NN method. Compared with other commonly-used algorithms, this computation method has higher localization when fault rate is low and can maintain better stability when fault rate is high.
Keywords
fault localization; execution trace; similarity degree; suspiciousness ratio; test caseText
DOI
10.12783/dtetr/iceta2017/19923
10.12783/dtetr/iceta2017/19923
Refbacks
- There are currently no refbacks.