A social graph based text mining framework for chat log investigation

摘要: This paper presents a unified social graph based text mining framework to identify digital evidences from chat logs data. It considers both users' conversation and interaction data in group-chats discover overlapping interests their ties. The proposed applies n-gram technique association with self-customized hyperlink-induced topic search (HITS) algorithm key-terms representing interests, key-users, key-sessions. We propose generation model interactions, where ties (edges) between pair of users (nodes) are established only if they participate at least one common group-chat session, weights assigned the on degree overlap interactions. Finally, we present three possible cyber-crime investigation scenarios user-group identification method for each them. our experimental results set comprising 1100 11,143 sessions continued over period 29 months January 2010 May 2012. Experimental suggest that is able key-terms, key-sessions, user-groups data, all which crucial investigation. Though recovered single computer, it very likely collected multiple computers real scenario. In this case, can be combined together generate more enriched graph. However, experiments show objectives achieved even computer by using draw relationships every users.