Variable heuristicsConst

Type declaration

• chi_square: ChiSquare

  Chi-square: classical statistical test for independence between term and class.

• gnd: GND

  GND (Google Normalized Distance): favors focused and discriminative terms.

• jlh: JLH

  JLH score: emphasizes sharp increases in frequency within the subset.

• mutual_information: MutualInformation

  Mutual information: captures all co-occurrence relationships between term and class.

• percentage: PercentageScore

  Percentage score: measures the proportion of term occurrences that fall within the subset. Can be overly sensitive to low-frequency terms

• solr: SolrRelatedness

  Used by Solr for the "relatedness" aggregation