Package org.terrier.matching.models

Provides the classes that implement various weighting models. Generally, the models fall into two classes:

• Term weighting models - score an occurrence of a document regardless of where the occurrences appear.
• Field-based weighting models - score an occurrence of a document depending on the fields that the term occurs in.

Various different families of weighting models are implemented:

• IDF: e.g. TF_IDF (aka BM11), BM25, LemurTF_IDF.
• Divergence from Randomness: e.g. PL2, InL2. Arbitrary DFR models can also be generated using DFRWeightingModel
• hyper-geometric Divergence from Randomness: e.g. DLH, DPH, DFRee
• Language modelling: e.g. Hiemstra_LM, Dirichlet_LM
• Divergence from Independence: DFI0

Of the field-based models, the following are implemented:

• BM25F
• PL2F
• MDL2
• ML2
• Arbitrary per-field normalisation models can be generated using the PerFieldNormWeightingModel model.

Class Summary

Class	Description
BB2	This class implements the BB2 weighting model.
BM25	This class implements the Okapi BM25 weighting model.
BM25F	A convenience subclass of PerFieldNormWeightingModel setup to do specifically BM25F, as described by [Zaragoza TREC-2004].
CoordinateMatch	A weighting model that returns 1 for each match.
DFIC	Divergence From Independence model based on Chi-square statistics (i.e., standardized Chi-squared distance from independence in term frequency tf).
DFIZ	Divergence From Independence model based on Standardization (i.e., standardized distance from independence in term frequency tf).
DFR_BM25	This class implements the DFR_BM25 weighting model.
DFRee	This class implements the DFRee weighting model.
DFReeKLIM	This class implements the DFReeKLIM weighting model.
DFRWeightingModel	This class implements a modular Divergence from Randomness weighting model.
DirichletLM	Bayesian smoothing with Dirichlet Prior.
Dl	This class implements a simple document length weighting model.
DLH	This class implements the DLH weighting model.
DLH13	This class implements the DLH13 weighting model.
DPH	This class implements the DPH hypergeometric weighting model.
Hiemstra_LM	This class implements the Hiemstra LM weighting model.
Idf	This class computes the idf values for specific terms in the collection.
IFB2	This class implements the IFB2 weighting model.
In_expB2	This class implements the In_expB2 weighting model, namely Inverse Expected Document Frequency model with Bernoulli after-effect and normalisation 2.
In_expC2	This class implements the In_expC2 weighting model.
InB2	This class implements the InB2 weighting model, namely Inverse Document Frequency model with Bernoulli after-effect and normalisation 2.
InL2	This class implements the InL2 weighting model.
Js_KLs	This class implements the Js_KLs weighting model, which is the product of two measures: the Jefrreys' divergence with the Kullback Leibler's divergence.
LemurTF_IDF	This class implements the TF_IDF weighting model as it is implemented in Lemur.
LGD	This class implements the LGD weighting model.
MDL2	This class implements the MDL2 field-based weighting model.
ML2	This class implements the ML2 field-based weighting model.
Null	A weighting model that returns 0 for each match.
PerFieldNormWeightingModel	A class for generating arbitrary per-field normalisation models.
PL2	This class implements the PL2 weighting model.
PL2F	A convenience subclass of PerFieldNormWeightingModel setup to do specifically PL2F.
SingleFieldModel	Use a normal weighting model on a pre-determine subset of the field.
StaticFeature	Class for query independent features loaded from file.
StaticScoreModifierWeightingModel	Base abstract class for query independent features loaded from file.
Tf	This class implements a simple Tf weighting model.
TF_IDF	This class implements the TF_IDF weighting model.
WeightingModel	This class should be extended by the classes used for weighting terms and documents.
WeightingModelFactory	A factory method for handling the initialisation of weighting models.
WeightingModelLibrary	A library of tf normalizations for weighting models such as the pivoted length normalization described in Singhal et al., 1996.
XSqrA_M	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.