Software Vulnerability Classification Based On Deep Neural Network
Markad Ashok Vitthalrao1, Mukesh Kumar Gupta2

1Markad Ashok Vitthalrao*, Research Scholar, Department of Computer Science & Engineering, Technology, Suresh Gyan Vihar University, Jagatpura, Jaipur, India.
2Dr. Mukesh Kumar Gupta, Principle, Department of Computer Science & Engineering, Technology, Suresh Gyan Vihar University, Jagatpura, Jaipur, India.
Manuscript received on September 16, 2019. | Revised Manuscript received on October 05, 2019. | Manuscript published on October 30, 2019. | PP: 3146-3150 | Volume-9 Issue-1, October 2019 | Retrieval Number: A9746109119/2019©BEIESP | DOI: 10.35940/ijeat.A9746.109119
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: Software vulnerability is most common issues in software engineering, many applications has suffering vulnerability, information leakage, and data hijacking such kind of problems facing since couple of years. Sometimes developers should be making some mistakes during code making which generate vulnerability issues for entire application. In this research work, we carried out an approach to software vulnerability detection using deep learning approach behalf of metadata processing. The system carried software vulnerability detection based on the Deep Neural Network (DNN). a new dynamic vulnerability classification approach has suggested. The model basic build based on TF-IDF as well density based feature selection approach for DNN. basically TF-IDF has used to measured the frequency and weight of specific word of vulnerability description; the Vector Space Model (VSM) is used for feature selection to achieve an finest set of feature term, and; the DNN neural network model is used to built an dynamic weakness classifier to achieve effectiveness into the bug detection. The overall system has categorized into four phases in first phase we detect the code clone to eliminate the data redundancy and execution time complexity, in second we apply Vector Space Model (VSM) recommend the re-factor possibility in entire code while in third section we build DNN module for software vulnerability detection and finally recommend the vulnerability for entire code. The system partial implementation has evaluated in java environment which provide satisfactory results for heterogeneous code modules.
Keywords: Deep neural networks, Computer security, Data mining, Machine learning.