期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2016
卷号：90
期号：1
出版社：Journal of Theoretical and Applied
摘要：This paper proposes application of GA-RBF Neural Network Algorithm in context of software fault localization. A neural network is trained on the basis of Code Coverage information of a test case and the corresponding execution result, successful execution or failure. The weights and structure of the RBF neural network is then optimized using Genetic algorithm. The hidden layer neurons number and connection weights are encoded using binary encoding and real encoding respectively. For further leaning, LMS (Least Mean Square) algorithm is used. A set of virtual test cases (each covering a single statement) is then given input to the trained and optimized network. The output of the network is considered to be �degree of suspiciousness� of the corresponding statement. Finally the statements are ranked on the basis of their corresponding degree of suspiciousness.
