An Enhanced Approach for Software Bug Localization using Map Reduce Technique based Apriori (MRTBA) Algorithm

Background/Objectives: Software bugs are generally the faults or errors that occur in the code that may leads to incorrect results. It is therefore necessary to find software bugs to increase the quality of software and for making the software to meet the user requirements. Methods/Statistical Analysis: The testing enables assessment of software which ensures the system whether it meets the system requirements. Graph mining is an approach to find software bugs and furthermore testing involves more computational complexity and cost therefore tools are developed for testing the software. The challenging task is to locate and fix bugs automatically. Bug localization using high performance map reducing process is very successful in the recent research. Findings: This proposed technique called Map Reduce Technique Based Apriori (MRTBA) Algorithm based on Graph mining in where edge weights are used to takes the program as the input and computes method calls of the program. Here each node represents method and each edge represents a method call. The algorithm aims for detection of faulty nodes. It uses Hash Map for reducing edge weights. JUnit test cases are performed at last for detecting bugs by giving different inputs. The approach mines all nodes not only at same level but also the nodes present at different levels. Conclusion/Improvements: In this work MRTBA, a dynamic control flow centered approach for bug localization has been presented. In future this technique can be extended with the proposal of steps involves in connecting same nodes at different levels. It also further enhances to find out indirect recursions for both the trivial and non-trivial map reduction in software bug localization.