Itemset Mining Based Episode Profiling of Terrorist Attacks Using Weighted Ontology

Itemset mining is a prominent research problem of data mining domain. It is extensively utilized for knowledge discovery in domains dealing with multi-component records or itemsets. Terrorism is a similar domain where every terrorist attack carries attack attributes such as location, target and attack type as components. Treatment of terrorist attack episodes as itemsets can facilitate effective pattern analysis and forecast of future attack episodes. This paper introduces a novel approach of mapping three major attributes of terrorist attacks taken place in a region in a single weighted ontology. The weighted ontology is later employed to discover and forecast useful information about possible attack episodes in the future.