An Empirical Study On Fault Localization And Effective Test Case Selection By Neural Network

A Radial basis function (RBF) neural network based fault localization technique is proposed in this paper to assist programmers in locating bugs effectively. Here we employ a three-layered feed forward artificial neural network with a radial basis function for its hidden unit activation and for linear function with its output layer activation. Here the neural network is trained to have a good relationship between the statement coverage information of a test case and its corresponding execution result to get a success or failure. The trained network is then given as an input to a set of virtual test cases, each covering a single statement, and the output of the network, for each virtual test case, is considered to be the suspiciousness of the corresponding covered statement. A statement with a higher suspiciousness has a higher likelihood of contain a bug, and thus, statement can be ranked in descending order of their suspiciousness. The Ranking can then be examined one by one, starting from the top, until a bug is located. Six case studies on different programs were conduced, with each faulty version contain a distinct bug, and the result clearly show that our proposed technique is much more effective than Tarantula, another popular fault localization technique.