Link Analysis of Knowledge Derived from Social Media Sources

ABSTRACT: Warfare is being transformed from a closed, state-sponsored affair to distributed terrorist activities enabled by social media. In Phase I of this research, Securboration developed the prototype Link Analysis Knowledge Environment (LAKE) to enable analyst driven social network analysis. LAKE Phase I demonstrated noise reduction, link analysis, and influential node identification within massive social network datasets. Our approach leverages Securborations existing suite of semantics-based content analysis software and research from renowned machine learning expert Dr. Carlos Guestrin and his associated GraphLab parallel processing toolkit. The result in LAKE is incomparable speed and accuracy in automatically analyzing complex social networks to identify linkages hidden deep within social media data streams. LAKE has also garnered the attention of the Combined Information Data exchange Network (CIDNE) community and the LAKE Team has attained advocacy for LAKE integration into CIDNE. Another transition path for LAKE has been initiated with the US Marshals Service to support executing outstanding arrest warrants. Additionally, the Joint Inter Agency Task Force South has expressed an interest in LAKE for their counter narcotics mission. In this Phase II Proposal, the LAKE Team details the objectives, workplan, and commercialization strategy to ensure long-term success and transition of this research. BENEFIT: LAKE is a force multiplier for the intelligence community; a Cloud based mechanism to add analytical capabilities quickly without investing in new infrastructure, training new personnel, or licensing new software. LAKE executes sophisticated learning and noise reduction algorithms within a high performance parallelized computing framework. LAKE will provide significant benefits to both the public and private sector. In the public sector, the need to understand complex social networks is highlighted on a near daily basis, to preempt a terror event, or understand the geo-political situation within a dynamic environment (e.g. Arab Spring). In both the public and private medical domain, link analysis can prove vital in combatting epidemics and pandemics by identifying social communities that may be exposed. In the private sector, there is significant opportunity for LAKE in the area of business intelligence. For example, Competitive Link Analysis enables a business to understand why people are linking to their competitors. LAKE can help businesses improve customer retention, cross-sell and up-sell by enabling marketers to identify social communities based on relationships between customers; measure and segment customers based on social influence; and target customers based on changes within their social networks.