ArangoSearch Scorers

ArangoSearch Scorers are special functions that allow to sort documents from a view by their score regarding the analyzed fields.

Details about their usage in AQL can be found in the ArangoSearch SORT section.

• BM25: order results based on the BM25 algorithm

• TFIDF: order results based on the TFIDF algorithm

BM25() - Best Matching 25 Algorithm

IResearch provides a ‘bm25’ scorer implementing the BM25 algorithm. Optionally, free parameters k and b of the algorithm typically using for advanced optimization can be specified as floating point numbers.

BM25(doc, k, b)

• doc (document): must be emitted by FOR doc IN someView

• k (number, optional): term frequency, the default is 1.2. k calibrates the text term frequency scaling. A k value of 0 corresponds to a binary model (no term frequency), and a large value corresponds to using raw term frequency.

• b (number, optional): determines the scaling by the total text length, the default is 0.75. b determines the scaling by the total text length.

  • b = 1 corresponds to fully scaling the term weight by the total text length
  • b = 0 corresponds to no length normalization.

At the extreme values of the coefficient b, BM25 turns into the ranking functions known as BM11 (for b = 1) and BM15 (for b = 0).

TFIDF() - Term Frequency – Inverse Document Frequency Algorithm

Sorts documents using the term frequency–inverse document frequency algorithm.

TFIDF(doc, withNorms)

• doc (document): must be emitted by FOR doc IN someView
• withNorms (bool, optional): specifying whether norms should be used via with-norms, the default is false