Link Analysis of Knowledge Derived from Social Media Sources

ABSTRACT: The internet has become an important part of the world"s communication networks. It has also become a resource for terrorists and criminals. Today, analysts must laboriously sift through vast amounts of data in order to"connect the dots". In this project, we propose to develop a scalable system for link analysis of data extracted from social media. Our work considers the process of knowledge extraction, information integration and analysis. The goal is to design an integrated framework that will enable Air Force intelligence analysts to efficiently analyze data from online sources. Probabilistic inference methods for classification, entity resolution and clustering can semi-automatically detect and link multiple aliases, document timelines, forensically analyze content, and assist analysts in identifying threats. BENEFIT: The Web has made it easy to access an enormous amount of open source data. The richness and size of this data has made it a priority to develop automated analysis methods that can help analysts predict, understand, and explain events around the world. This project will produce a scalable solution that will help analysts make sense of, and monitor, events being discussed or planned online. Analysts want to understand who the protagonists are, what they have done, and what they are planning. They have limited time and manpower available to do this by hand, but they also have expertise and commonsense that is far beyond the present capabilities of software to replicate. Thus, our approach is to develop semi-automated methods that analysts can intuitively interact with in order to do their job more efficiently. The software will not only be useful for government defense and intelligence analysts, but will also address the needs of the commercial background investigation industry.