PREDICT SOFTWARE VULNERABILITY

Computer Science And Engineering

Changes made in para- Security concerns are primarily brought about by software defects. If a hostile attack exploits a weakness, the system's safety will be gravely compromised. Even catastrophic losses could result. Automatic classification techniques are thus advantageous for managing software vulnerabilities effectively and enhancing system security. It will lessen the possibility of system compromise and attack. Model for automatically classifying vulnerabilities (IGTF-DNN) In this study, a novel model based on term frequency-deep neural networks dubbed Information Gain has been put forth. Deep neural networks (DNN) and information gain (IG), based on frequency- inverse document frequency, are used to build the model (TF-IDF). The frequency and weight of phrases extracted from vulnerability descriptions using TF-IDF are determined using Information Gain, and the optimal set of feature words is assembled using Choose features. The automatic vulnerability classifier is then built utilising a deep neural network model to categorise vulnerabilities effectively. The effectiveness of the suggested model has been evaluated using the US National Vulnerability Database. The TFI- DNN model performs better on assessment indices like precision and recall measures when compared to KNN. It may not only improve the effectiveness of the vulnerability recovery and management, but also reduce the danger of systems being attacked and collapsing, which is critical for systems' security capabilities, assuming the vulnerability can be classified and handled with effectiveness. A growing number of studies on vulnerability classification are being undertaken by qualified security researchers as software security vulnerabilities play is a significant part in cyber-security assaults.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.