Research on Methods of Finding Suspicious Patterns about Programs

LIU Xuemei,GU Guochang,LIU Yongpo,WU Ji,JIA Xiaoxia

DOI: https://doi.org/10.3969/j.issn.1006-7043.2011.02.013

2011-01-01

Abstract:Due to the dynamic execution of object-oriented programs,the object states cannot be described and analyzed properly by use of traditional test methods.Doing so will affect the accuracy of fault localization in the object-oriented programs.On the basis of the concept of a pattern,the lifetime behavior model(LBM) was proposed to describe the behavioral characteristics of objects by using sequence patterns.Also,as the behavioral patterns of a class were constituted of LBM,the behavioral patterns of a program were defined on the basis of the behavioral patterns of classes.Next,the framework of finding suspicious patterns in a program was designed;this framework involved preparing data,extracting patterns,and finding abnormal behaviors.The standard of extracting patterns was defined and the mining algorithms of finding patterns were described.Lastly,the methods of finding suspicious patterns whose length was equal and larger than 2 were put forward.These methods can be utilized in research on fault localization,and they gain the rate of 0.761 in precision and 0.737 in recall on average as indicated by the experimental data.Satisfactory results have been achieved with these methods.