By Nick Heath, 23 September 2008 14:54
Key people, places and objects - known as entities - from the reports are listed and can be used to refine future searches.
Reports in any language can be searched thanks to the system's ability to eliminate "background language" and only list entities. It does not understand the language itself but determines whether a word is a proper noun using statistical analysis of how frequently it appears in the reports.
Here, it is obvious that the Abu Sayyaf group (ASG) and the Jemaah Islamiya Organisation (JI) are active in the Philippines.
These names can then be fed into fresh queries to uncover what other people and places are associated with the groups.
Only significant entities as flagged up by the system, as each entity is assigned a rank, giving common nouns such as Manila a low score, while rating groups like the ASG highly.
Photo credit: Detica



In order to post a comment you need to be registered and logged in.
Log in or create your silicon.com account below