Information retrieval modules

Information retrieval modules provide comprehensive information retrieval functionality including:

Indexer module

Indexer module creates an updatable, multiple-field full-text and metadata index.

Indexing can be performed on document content and metadata as well as on results of other analysis methods (such as information extraction). Supported data types include text, numerical, and date/time, whereby the last two can be supported with arbitrary precision. A special case is indexing of hierarchical or graph structures where the neighbourhood of each node is also considered.

Search module

Search module executes search queries against the indexed information to select relevant document sets.

Provided are comprehensive document searching capabilities including:

  • full text search (case sensitive and insensitive, stemmed and non-stemmed)
  • metadata search
  • range search
  • wildcard search
  • fuzzy search
  • Boolean queries
  • search by example (also with considering structure information and multiple examples)
  • concept search
  • relevance feedback
  • query expansion
  • suggestions
  • etc.

Hits are retuned as a ranked list sorted by relevance including metadata and content snippets. Faceted search capability provides filtering over aggregated metadata-fields, such as for examples over people or geographic locations.