Modeling of a Software Vulnerability Identification Method

Software vulnerabilities are becoming more and more increasing, their role is to harm the computer systems of companies, governmental organizations and agencies. The main objective of this paper is to propose a method that will cluster future software vulnerabilities that may spread. This method is developed by combining the Multiple Correspondence Analysis (MCA), the Elbow procedure and the Kmeans Algorithm. A simulation was done on a dataset of 15713 observations. This simulation allowed us to identify families of future vulnerabilities. This model was evaluated using the silhouette index.