A Similar Code Analysis Techniq Ue Based On Diversified Program Characterstics

With the rapid development of information technology,software has reached an unprecedented scale.Similar code analysis technology plays an important role in program plagiarism detection,intellectual property protection,information search,and malicious code monitoring.It has received extensive attention from the academic community.However,the detection method based on attribute counting cannot fully represent the grammatical and semantic features of the program,so its accuracy is low.The detection method based on structure measurement is complex and inefficient.The existing similar code analysis tools are not flexible enough to carry out multilevel detection.It is quite necessary for us to solve the above problems.This paper presents a similar code analysis technique based on diversified program characterstics.This paper is based on the topic—research and implementation of a diversified characterstics extraction approach for program similarity analysis.This paper first defines the concept of code similarity and divides it into three situations—simple similarity,structural similarity and functional similarity.For these three cases,this paper make a detailed analysis from four levels,which are structured statement block,function,file and program.Then this paper develops different analysis strategies for different levels.The process of high-level analysis relies on the results of low-level analysis.This method not only focuses on the macro information of the code but also the internal information of the code,improving the efficiency of analysis and achieving higher accuracy.Based on the above methods,this paper implements a similar code analysis system This paper uses four open source projects for experiments.Through the analysis of the experimental results,we come to the following conclusions:This method can improve analysis efficiency and has high accuracy.