Research On Behavior Pattern Mining Algorithm In Software Dynamic Execution

The scale and complexity of computer software are increasing,and it produces a large number of software execution tracks every day,which makes the software security face a series of challenges.Therefore,the analysis of complex software systems has increasingly become a hot issue in data mining.In this paper,the sequential pattern mining algorithm is used to mine the software's critical path pattern from a large number of software execution tracks and analyze the software behavior.The main work is as follows.Firstly,in view of the variety of functions and the behavioral complexity of software dynamic execution process,the dynamic execution traces of software are tracked.It can obtain the calling state and calling sequence of the function,and then the software sequence model is constructed.Besides,the data is abstracted to get the function call graph.Based on the complicated function call graphs,a method to measure the probability of failure of a node is designed.Secondly,aiming at the problem of fault finding in the software execution process,it is difficult to find the source of the fault.In this paper,we design a high utility path pattern mining algorithm Huspm in software execution sequence.It takes the sub sequence pattern as a unit to consider the pattern utility,which is more in line with the orderliness and continuity of software execution tracks.It calculates the importance of nodes and defines their internal utility values,according to the different dependencies among the nodes in the software network.Based on pattern growth,the high utility path patterns in software execution trace are discovered,and the location of the fault source can be determined quickly and accurately.Again,in order to analyze the process of fault propagation in software execution,a maximal high utility sequential pattern mining algorithm MHuspm is ProPosed.It constructs software execution sequence data into a HUSP-Tree tree structure.Based on the HUSP-Tree structure mining the maximal high utility path patterns,which is of great significance for the analysis of the fault propagation process in software execution trace.Finally,we verify the ProPosed algorithms.The algorithms ProPosed in this paper are all implemented in the Java language and run on Windows systems.Through experimental analysis,the algorithms of this paper are evaluated in terms of performance such as runtime,memory consumption,and scalability.