مقالے کی معلومات
شمارہ
مقالے کی قسم
زبان
تلخیص
A vulnerability discovery model (VDMs) play a central role to model the rate at which vulnerabilities are discovered for software. Though, these models have various shortcomings viz., multi VDMs, changes in VDMs, and development of new VDMS for different datasets due to diverse approaches and assumptions in their analytical formation. There is a clear need for intensive investigation and extensive use of these models to enhance the predictive accuracy of existing VDMs. In this paper, to enhance the predictive accuracy of existing VDMs, a multiple error iterative analysis method (MEIAM) along with artificial neural network sign estimators has been proposed based on the residual errors. Our findings reveal that the proposed method optimizes to fit historical vulnerability accurately and helps to predict future trends of vulnerabilities across different datasets and models. Repeated calculations of residual errors using these models are used to improve and adjust the forecast accuracy to the expected level. The experiment performed by using real vulnerability data of three type’s popular software: Windows 10 (613), Android 7.0 (1018), Internet Explorer 11 (60), and Firefox 20 (502), starting from the first day of the issue or the earliest available in NVD database. The results demonstrate that the method is universally applicable to any of the VDMs to improve predictive accuracy.
Gul Jabeen1,2, Sabit Rahim1, Gul Sahar1, Akber Aman Shah2, Tehmina Bib. (2020) An Optimization of Vulnerability Discovery Models using Multiple Errors Iterative Analysis Method, , PROC. OF THE PAS: A; 57, Issue 3.
-
Views
243 -
Downloads
41