Extended Framework for Target Oriented Network Intelligence Collection

摘要: The Target Oriented Network Intelligence Collection (TONIC) problem is the of finding profiles in a social network that contain publicly available information about given target profile via automated crawling. Such are called leads. Leads can be found by crawling using profiles' friend lists (immediate neighborhood) order to decide which will crawled next. Assuming leads tend cluster together, prior work limited search for new only immediate neighbors previously found. In this paper we relax limitation, and extend scope wider neighborhood, including possibility non-leads, i.e., have no target. We propose set heuristics guide search. Experimental results show with setting more faster. addition, perform cost benefit analysis search, weighing reward costs