Framework for surveillance of instant messages

Instant messengers (IMs) and social networking sites (SNS) such as Facebook may contain harmful and suspicious messages, which is of national security concerns. Organised crimes have adopted online chatting technique to send these suspicious messages as these systems have all the facilities and could serve as platform to spread across their information widely through socio-engineered and general text messages. A solution to this problem is to detect suspicious messages from the typed messages. In this paper, we proposed a suspicious message detection system (SMDs) to detect suspicious messages. SMDs framework makes use of databases where instant messages are stored and an ontology information extraction technique which is able to detect suspicious messages using probabilistic models. The objective of SMDs framework is to trace the identified criminals by browsing their profile details available from their e-mail account, where suspicious messages are discovered during online chat. Experimental analysis is evaluated using the user generated content (UGC) testbed which consist of suspicious messages for eight different test cases with user-defined threshold value tested using SMDs. The results obtained shows high precision rate compared to the existing state-of-the-art systems.