The vulnerability exposure of interconnected device in the cyberspace keep surging with evolving technology. Gathering social and technical indicators useful for threat intelligence from open-source data is essential in developing preventive measures and strategies to prevent cyberattack before actual occurrence or mitigate the effect of vulnerability exposure. This research developed a novel framework for identifying existing threats and predicting potential vulnerability exposure with our trained model. The methodology used in this research showcased that potential cyberthreat can be predicted from open-source data using a deep learning algorithm(LSTM). The developed model achieved accuracy of 91%, precision of 90% and recall of 91% on test data. We utilized open-source intelligence to identify existing threat and the severity level of cyberattack by crawling the National vulnerability Database(NVD) and Common Vulnerabilities and Exposures (CVE) Database for a list of publicly known cyberthreats related to the input search predicted as threats by our model.
Authors: Victor Adewopo, Bilal Gonen, Nelly Elsayed, Sylvia Azumah