An introduction to framework of security event capturing and tracking

Nowadays, microblogs provide abundant text data. Mining valuable information from it becomes an interesting topic. Based on the topic model, things can be easily traced. The complete view of an event can help people make decisions. Once we can track topics in a specific domain, experts can rapidly react according to important events. In this paper, we proposed a LDA [2] based framework to capture emergent events in the information security domain from Twitter data. Recently, different types of malwares, bugs and phishing attacks are detected on the internet. Predicting the novel attacking technique can help users to prevent infected. According to our system, we can track an event to realize when it starts, how it changes and the event vanishes. The real-time intelligence we extract makes not only users but also service providers rapidly react. Different from state-of-art topic models, we focus on topic changement with time pass.