org.apache.lucene.search.similarities

Class IBSimilarity



  • public class IBSimilarity
    extends SimilarityBase
    Provides a framework for the family of information-based models, as described in Stéphane Clinchant and Eric Gaussier. 2010. Information-based models for ad hoc IR. In Proceeding of the 33rd international ACM SIGIR conference on Research and development in information retrieval (SIGIR '10). ACM, New York, NY, USA, 234-241.

    The retrieval function is of the form RSV(q, d) = ∑ -xqw log Prob(Xw ≥ tdw | λw), where

    • xqw is the query boost;
    • Xw is a random variable that counts the occurrences of word w;
    • tdw is the normalized term frequency;
    • λw is a parameter.

    The framework described in the paper has many similarities to the DFR framework (see DFRSimilarity). It is possible that the two Similarities will be merged at one point.

    To construct an IBSimilarity, you must specify the implementations for all three components of the Information-Based model.

    1. Distribution: Probabilistic distribution used to model term occurrence
    2. Lambda: λw parameter of the probability distribution
      • LambdaDF: Nw/N or average number of documents where w occurs
      • LambdaTTF: Fw/N or average number of occurrences of w in the collection
    3. Normalization: Term frequency normalization
      Any supported DFR normalization (listed in DFRSimilarity)
    See Also:
    DFRSimilarity
    • Constructor Detail

      • IBSimilarity

        public IBSimilarity(Distribution distribution,
                            Lambda lambda,
                            Normalization normalization)
        Creates IBSimilarity from the three components.

        Note that null values are not allowed: if you want no normalization, instead pass Normalization.NoNormalization.

        Parameters:
        distribution - probabilistic distribution modeling term occurrence
        lambda - distribution's λw parameter
        normalization - term frequency normalization
    • Method Detail

      • score

        protected float score(BasicStats stats,
                              float freq,
                              float docLen)
        Description copied from class: SimilarityBase
        Scores the document doc.

        Subclasses must apply their scoring formula in this class.

        Specified by:
        score in class SimilarityBase
        Parameters:
        stats - the corpus level statistics.
        freq - the term frequency.
        docLen - the document length.
        Returns:
        the score.
      • explain

        protected void explain(List<Explanation> subs,
                               BasicStats stats,
                               int doc,
                               float freq,
                               float docLen)
        Description copied from class: SimilarityBase
        Subclasses should implement this method to explain the score. expl already contains the score, the name of the class and the doc id, as well as the term frequency and its explanation; subclasses can add additional clauses to explain details of their scoring formulae.

        The default implementation does nothing.

        Overrides:
        explain in class SimilarityBase
        Parameters:
        subs - the list of details of the explanation to extend
        stats - the corpus level statistics.
        doc - the document id.
        freq - the term frequency.
        docLen - the document length.
      • toString

        public String toString()
        The name of IB methods follow the pattern IB <distribution> <lambda><normalization>. The name of the distribution is the same as in the original paper; for the names of lambda parameters, refer to the javadoc of the Lambda classes.
        Specified by:
        toString in class SimilarityBase
      • getLambda

        public Lambda getLambda()
        Returns the distribution's lambda parameter