Test Case Minimization for Object Oriented Testing Using Random Forest Algorithm

Abstract

Software maintenance is one of the most costly and crucial phases in the life cycle of software. It consumes almost 70% of the resources and cost of the software. Software testing aims to execute or examine the software with the intention of detecting the faults in it. Reducing the cost of the testing process is one of the major concerns of the testers. With the growing complexities in Object Oriented (OO) software, the number of faults present in the software module is increased. In this paper, a technique has been presented for minimizing the test cases for the OO systems. A case study of Xerces 1.4 open source software is carried for the evaluation of proposed technique. The mathematical model used in the proposed methodology was generated using the open source software WEKA. The approach is based on selecting significant Object Oriented metrics. Highly Efficient, Less efficient or inefficient Object Oriented metrics were identified by the techniques based on feature selection. Test case generation and minimization is achieved on the basis of coverage of highly fault prone classes. To minimize the test cases, proposed methodology used only significant OO metrics for assigning weights to the test paths. The proposed work promisingly reduced the cost and time taken during test suite minimization.