Algorithm Research Of Software Behavior Characteristics From Software Execution Graph

In the era of big data,the scale and application fields of software system is becoming more and more complex,which makes software dynamic execution path to become more decision-making and diversity of information assets with a novel model.Therefore,it becomes basic guarantee of quality of software development and can help to reasonable application and development in the field of applied software,which including the effectively mining of the inherent characteristics of the software,the realization of the behavior feature modeling and the research of the algorithm based on software dynamic execution path.This topic excavates behavior patterns of interest to the user from large software execution trace,in order to apply to a wide range of software security analysis,software quality evaluation,software vulnerabilities,and other areas of the application software system.Firstly,this paper proposes a method of building software execution graph by using software dynamic execution path.Based on different granularity of software,the unit weight and the utility values of sequence are generated by the distribution function under different granularity.The depth-first traversal method was used to construct the software execution sequence,which consider each sequence as a software execution sequence of functions,and the function call sequence database are formed in the end.Secondly,an approach for mining promising extension-based high utility patterns from software function call graph is proposed.The algorithm translates the software dynamic execution path into a sequence,and designs the table structure which is used to store the utility and index information of the pattern and called Pre-Utility-List.To store the utility values of the patterns whose length is 2,a novel structure named PEUS(Promising Extension Utility-based Structure)was proposed,and a pruning strategy of potential high utility pattern is raised based on this structure,which can decrease the upper bound of pattern utility.Thirdly,a novel algorithm for mining top-k high utility traversal patterns from software dynamic executing graph is presented.Based on the functions call order and dependencies in the software execution sequence,the strategy of Pre-insertion is put forward to improve the initial utility threshold.The usage of Sorting strategy and Filteringstrategy can help to expand potential high utility patterns early,and improve the efficiency of algorithm.Finally,the algorithm proposed in this paper are implemented on the platform of Windows in Java language.The run time,the number of candidates and the memory usage performance of the space are analyzed by the experiment.