Using a specific machine learning technique, this thesis proposes a method to identify suspicious statements during debugging which applies principles similar to Tarantula, an existing suspicious statement identification technique, but addresses its main flaw: its difficulty of dealing with the presence of multiple faults since it assumes that failing test cases execute the same fault(s). The improvement we present in this thesis is through the use of C4.5 decision trees to identify various failure conditions based on information regarding the test cases' inputs and outputs. Failing test cases executing under similar conditions are then assumed to fail due to the same fault. Statements are then considered suspicious if they are covered by a large proportion of failing test cases that execute under similar conditions. We report on two case studies that demonstrate improvement over the original Tarantula technique in terms of statement ranking. Another contribution of the thesis is to show that failure conditions as modeled by a C4.5 decision tree accurately predict failures and can therefore be used to help debugging.
展开▼
