Design Code Clone Detection System uses Optimal and Intelligence Technique based on Software Engineering

Abstract

Code clones are the main source of cloned software. Now, a day’s redundancy in initial code is called clones or duplicate code caused by copy and paste, could search consistently using code clone detection software tools. The redundancy could arise also individually, although, not produced by copy and paste. Recently, it is not clear to define how the only Metric approach (functionally same clones) dissimilar from duplicates made by copy and paste. In this paper, our idea is to understand and classify the syntactical dissimilates in (function same clones) Metric based technique used with the help of swarm and artificial intelligence techniques that described them from copy and paste code clones in a path that helps clone detection research. In this method, we discussed it by functionally using the same program in Java, C++, and MATLAB for coding challenges. We studied syntactic correspondence with new detection software tools and discovered whether code clone detection can perform outside other structure. We implement the metric based approach extract the code properties i.e. LOC, Function Overloading, Function Repetition, Total number of functions, Global and Local Variable with the help of PDG and AST tree code clone techniques. The Classification with Neural Network approach to classified the code clone and calculates percentage of the code clone as compared to original code. We executed all tools on 100 programs and manually classified the dissimilarity in a random code sample of 60 programs files. We search non-metric function similar codes, where complete records were syntax and syntactically same code. The major difference between metric and crossbreed algorithm code clone detection techniques are beyond the recent code clone detection approaches.