Detection of Metrics Based Code Cloning Using Optimised SVM Algorithm

EasyChair Preprint 6641

Abstract

Code segments usually arises due to replication from one place and then rewrite them in to another segment of code with or without variations / changes are software cloning and the copied code is called clone. Generally Code Cloning defines the designing and development of software systems. Detection can be done based on Textual analysis, Lexical analysis, Syntax analysis, Se-mantic analysis, Hybrid analysis and Metric analysis. The major drawback of the research is that it emphasises more on fragments of copied code and does not focus on the feature that the fragments of duplicated code are may be part of a larger replicated program. In this process, many methods take a lot of time and it generates intricacy. In our study, a program source code is inspected for detecting various methods by adopting an “OPTIMIZED SVM ALGORITHM” and the method definitions are extracted and collected by means of a CLONE CODE. To evaluate the performance parameters we calculate the LOC, the number of recurrences, and maximum and minimum LOC. To enhance the performance metrics precision recall, accuracy and reduce the error rate and time complexity.

Keyphrases: Functional Clones, SVM, clone detection, fitness function.

@booklet{EasyChair:6641,
  author    = {S Karthik and B Rajdeepa},
  title     = {Detection of Metrics Based Code Cloning Using Optimised SVM Algorithm},
  howpublished = {EasyChair Preprint 6641},
  year      = {EasyChair, 2021}}