Corpus Level Fusion
After Semantex's natural language processing modules are complete the information in the tokenlist is stored in a proprietary index which preserves all analysis output in a corpus. This reflects an advancement to the level of state-of-the-art in information fusion, facilitating:
- Elimination of the need to reprocess documents if the ontology changes.
- Incorporation of existing structured data.
- The ability for users to interact with the system, including the addition of reference times which can be propagated throughout the corpus and further merging of entity profiles based on user specified aliases.
CORPUS-LEVEL ENTITY PROFILE GENERATION
Corpus-level entity profile generation enables rich entity summaries to be generated that consolidate mentions of an entity across a corpus. This includes entity relationships and any key events associated with the entity in question; leading to more intuitive browsing.
DYNAMIC EVENT RETRIEVAL
Dynamic event retrieval is motivated by the needs of intelligence analysts who frequently must search for new events based on changing situations or data sources.
Typically, event detection involves customizing the information extraction engine by including specific grammars to detect new events. Dynamic event retrieval provides a method for searching for new events without having to reprocess documents; it operates on a the stored, parsed corpus.
