CAR:Dictionary based software forensics method

With the software inheritance, reuse and piracy becoming more and more common, the software forensics has drawn much more attention increasingly; however, the traditional software similarity detection methods are mainly based on text, grammatical pattern or semantic analyses without deep mining software characteristics. With an attempt to better deal with these problems, this paper proposed the CAR, short for clustering, adjusting and rearranging method as to the software forensics. It focuses on computing the software similarity from the programming characteristics of source codes so as to assist in identifying the software authors by using the clustering algorithm. Experiments proved that whether the comparison of different software versions or the authorship identification of specific software show relativelyidealresults. Consequently, thisproposed method has practical value in the authorship disputes, proof of authorship in court and code re-engineering.