Software Vulnerability and Application Security Risk

Abstract

This research investigates the software vendor-based relationships between software vulnerability and application security risk. The data is obtained from the China National Vulnerability Database of Information Security (CNNVD). At first, we use the latent class model to classify the software vendors into three categories, and then employ regression models to estimate relationships between software vulnerability and application security risk for each of the three categories of the software vendors. The results show the relationships vary across the software vendors. The findings suggest that an IT vendor should learn specific vulnerability features according to its type to effectively avoid vulnerability generation on their products.