Techvar: Classification of Similarity in Software Detection Model using Deep Learning

In Industry 4.0, Deep Learning techniques have become an important research tool in many area namely healthcare, automobiles, video analysis, audio analytics, software systems etc. In recent years, many research are performed in software analysis using modern technologies. CrosLSim, SimMax, and atrpos models are used to review the software similarity detection techniques for different software systems. The existing models for similarity detection are not efficient to be used in major software projects. In this paper, the Techvar-DNN system which performs enhanced Probability, maintainability, testability, and reusability has been proposed. When compared with other methods namely Random Forest and Support vector machines, the proposed system provides increased recall, a smaller function size, and more efficient computing. Moreover, the proposed model results show better F-measure, precision and recall to improve the software similarity detection in a more efficient manner.