Class XSqrA_M

  • All Implemented Interfaces:, java.lang.Cloneable, Model

    public class XSqrA_M
    extends WeightingModel
    This class implements the XSqrA_M weighting model, which computed the inner product of Pearson's X^2 with the information growth computed with the multinomial M. It is an unsupervised DFR model of IR (free from parameters), which can be used on short or medium verbose queries.

    XSqrA_M has a high performance, and in particular has statistically significant better MAP performance than all other supervised models on the GOV2 collection. MAP for short (title only) and medium (title+description) topics, and comparative p values (two-tailed paired t-test) compared to supervised models (with optimal MAP parameter values) are as follows:

    QueriesMAP of XSqrA_MLGDDirichlet_LMPL2BM25In_expB2
    short0.3156 p=0.3277p=0.0075p=0.0055p=0.0064p=0.0002

    References Frequentist and Bayesian approach to Information Retrieval. G. Amati. In Proceedings of the 28th European Conference on IR Research (ECIR 2006). LNCS vol 3936, pages 13--24.

    Gianni Amati
    See Also:
    Serialized Form
    • Constructor Detail

      • XSqrA_M

        public XSqrA_M()
        A default constructor to make this model.
    • Method Detail

      • getInfo

        public final java.lang.String getInfo()
        Returns the name of the model, in this case "XSqrA_M"
        Specified by:
        getInfo in interface Model
        Specified by:
        getInfo in class WeightingModel
        the name of the model
      • score

        public final double score​(double tf,
                                  double docLength)
        Uses XSqrA_M to compute a weight for a term in a document.
        Specified by:
        score in class WeightingModel
        tf - The term frequency of the term in the document
        docLength - the document's length
        the score assigned to a document with the given tf and docLength, and other preset parameters