System “AntiTerror”
       The System provides a selection from natural language texts the relevant information (named entities) relating to terrorist groups and individuals.

          

      The main task of the System is the selection from the flow of media reports of those documents that relate to terrorist activity, with subsequent analysis of these documents and extraction of terrorist groups, organizations and individual terrorists, their status, place of residence or stay, as well as acts committed by them. As a result, automatically generate a knowledge base (KB), which served as the basis for semantic searching, responding to requests in a free form, the formation of hypotheses and expert decisions.
     Language processor (LP) has been devdloped in accordance with the peculiarities of the domain and tasks. In the LP additionally provides the extraction of following named entities:
• terrorist groups and organizations (Terrorizm);
• member of terrorist groups with indicating its role (leader, leader, etc.);
• the armed forces used to fight against terrorism (Military_Force);
• time’s intervals.
     Linguistic knowledge (LK) has been developed to select these entities. In accordance of the texts specifics Knowledge LZ was completed by new rools which provide the extraction of new entities, for example, the extraction of events places, expressed in forms "in the 25 km. from Kabul" or "camp near the city Umma," etc. Particular difficulties caused the extraction of Arab names and their constituent elements: abd (slave), Abu (father), Ibn or Ben (Son), etc. They strongly differ from the European standards, for example, Abd al-Rasul Ben Ahmad. For the known terrorists, as a rule, use the abbreviated names such as Ben Laden instead Usama Ben Laden etc.
     The system was designed for an integrated fight against terrorism.