SOFTWARE VULNERABILITY CLASSIFICATION MODEL USING NEURAL NETWORK
| Main Authors: | Ms. N. Zahira Jahan M.C.A., M.Phil, S. Madeshwaran |
|---|---|
| Format: | Article Journal |
| Bahasa: | eng |
| Terbitan: |
, 2022
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/6387693 |
Daftar Isi:
- Security risks are caused mainly due to software vulnerabilities. If any vulnerability is exploited due to a malicious attack, it will greatly compromise the system’s safety. It may even create catastrophic losses. So, automatic classification methods are enviable for effective management of vulnerability in software, thereby improving security performance of the system. It will reduce the risk of system being attacked and spoiled. In this study, a new model has been proposed named automatic vulnerability classification model (IGTF-DNN) Information Gain based on Term Frequency - Deep Neural Network. The model is built using information gain (IG) which is based on frequency-inverse document frequency (TF-IDF), and deep neural network (DNN): TF-IDF is used for calculating frequency/weight of words prepared from vulnerability description; Information Gain is used to select features for gathering optimal set of feature words. Then deep neural network model is used to construct an automatic vulnerability classifier to achieve effective vulnerability classification. The National Vulnerability Database of the United States has been used to test proposed model’s effectiveness. Compared to KNN, the TFI-DNN model has achieved better performance in evaluation indexes which includes precision and recall measures.