Generation of test oracles using neural network and decision tree model

Software Testing is a very important phase in the cycle of software development. It is the only phase which ensures the reliability on the software. Generally 40-50% of the software development cost is spent on this phase. Though many automatic testing tools are present, but still most research is required in this field to reduce cost and time allotted for this phase. Test Oracle is a process which ensures that expected Oracle is obtained for the given input which increase automation process. In search - based test generation techniques test inputs are generated in a large amount. To generate expected output for each and every test input is quite a difficult and time taking job. Test - Oracle is a mechanism that can determine the expected or predicted output for each given test input. In this paper, two approaches are discussed for generation of expected outputs with the help of artificial neural network and data mining approach i.e. decision tree.