Reproducible Analysis and Intelligent Scientific Criteria in Engineering Papers Classification using Data Science

A V S Pavan Kumar,Debrup Banerjee,Venkata Naresh Mandhala,Siva Koteswararao Chinnam

In this paper we discuss, any work must be reproducible in order to influence research and contribute to our profession's knowledge. Nevertheless, studies show that 70% of university laboratory work cannot reproduce. Reproducible work with not always available data sets or methods not clearly specified is uncommon in software engineering and complex specifications engineering. The lack of reproducible re-search prevents development, which means that researchers must replicate a scratch study. The RE researcher will read conference papers, find empirical articles and then review data that can (if available) be replicated in the empirical report. This paper deals with two aspects of the problem and discusses in RE articles RE documents and theoretical documents. In learning and development of an automatic classification in RE and empirical document identification, recent conference papers and RE documents have been used. They use the ERRC approach for performing supervised lecture classifications using natural language therapy and machine learning. Our software is equated with a fundamental keyword approach. From the study in all but a few cases, we found that the ERRC method worked better than the reference method.

Volume 12 | Issue 2

Pages: 1969-1972

DOI: 10.5373/JARDCS/V12I2/S20201242