---

XLingRetMethod Class Reference

#include <XLingRetMethod.hpp>

Inheritance diagram for XLingRetMethod:

List of all members.

Public Methods
	XLingRetMethod (const Index &dbIndex, const Index &background, PDict &dict, ScoreAccumulator &accumulator, double l, double b, bool cacheDR, string &sBM, string &tBM, const Stopper *stp=NULL, Stemmer *stm=NULL)
	Constructor.
virtual	~XLingRetMethod ()
	clean up.
virtual DocumentRep *	computeDocRep (int docID)
	Create a document representation.
virtual double	matchedTermWeight (int id, double weight, const DocInfo *info, const DocumentRep *dRep) const
	Score a given term for a given document.
virtual double	adjustedScore (double origScore, double pge) const
	Adjust the score for a given document.
virtual void	scoreCollection (const QueryRep &qry, IndexedRealVector &results)
	Score all documents in the collection.
virtual void	scoreInvertedIndex (const QueryRep &qryRep, IndexedRealVector &scores, bool scoreAll=false)
virtual QueryRep *	computeQueryRep (const Query &qry)
	compute the representation for a query, semantics defined by subclass
virtual QueryRep *	computeTargetKLRep (const QueryRep *qry)
virtual double	scoreDoc (const QueryRep &qry, int docID)
	Score a document identified by the id w.r.t. a query rep.
virtual void	updateQuery (QueryRep &qryRep, const DocIDSet &relDocs)
	update the query -- noop
Protected Methods
virtual double	scoreDocVector (const XLingQueryModel &qRep, int docID, FreqVector &docVector)
Protected Attributes
double	lambda
double	beta
double	numSource
double	numTarget
bool	docBasedSourceSmooth
bool	docBasedTargetSmooth
ScoreAccumulator &	scAcc
PDict &	dictionary
Stemmer *	stemmer
const Stopper *	stopper
const Index &	source
DocumentRep **	docReps
	cache document reps.
bool	cacheDocReps
	whether or not to cache document representations
int	docRepsSize
	number of documents plus 1, the size of the docReps array.
ScoreAccumulator *	termScores

---

Detailed Description

Cross lingual retrieval method. Translation dictionary based retrieval, scoring queries in the source language against documents in the target language using:
P(Q_s|D_t) = Prod_w_in_Q_s(lambda(Sum_t_in_D_t P(t|D_t)P(w|t) + (1-lambda)P(w|G_s)
where G_s is the background model for the source language.

---

Constructor & Destructor Documentation

XLingRetMethod::XLingRetMethod (  const Index &    dbIndex, const Index &    background, PDict &    dict, ScoreAccumulator &    accumulator, double    l, double    b, bool    cacheDR, string &    sBM, string &    tBM, const Stopper *    stp = NULL, Stemmer *    stm = NULL )

Constructor. Parameters: dbIndex  index for target language documents background  index for source language background model dict  PDict containing source->target translation probabilities accumulator  ScoreAccumulator for intermediate results l  lambda value to use for smoothing background model b  beta value to use for smoothing P(t|D) cacheDR  whether or not to cache document reps sBM  whether to use term frequency (tf/|V|) or term doc frequency (docCount(t)/Sum_w_in_V(docCount(w))) for the source language background model. Default is term frequency. tBM  whether to use term frequency (tf/|V|) or term doc frequency (docCount(t)/Sum_w_in_V(docCount(w))) for the targetlanguage background model. Default is term frequency. stp  source language Stopper to use when getting translations. stm  source language Stemmer to use when getting translations.

XLingRetMethod::~XLingRetMethod (    )  [virtual]

clean up.

---

Member Function Documentation

virtual double XLingRetMethod::adjustedScore (  double    origScore, double    pge )  const [inline, virtual]

Adjust the score for a given document. Parameters: origScore  the original score pge  the background probability to adjust by. Returns: log((lambda * origScore) + ((1 - lambda) * pge))

DocumentRep * XLingRetMethod::computeDocRep (  int    docID )  [virtual]

Create a document representation. Parameters: docID  the internal document id to create the representation for Returns: An instance of XLingDocRep

virtual QueryRep* XLingRetMethod::computeQueryRep (  const Query &    qry )  [inline, virtual]

compute the representation for a query, semantics defined by subclass Implements RetrievalMethod.

QueryRep * XLingRetMethod::computeTargetKLRep (  const QueryRep *    qry )  [virtual]

virtual double XLingRetMethod::matchedTermWeight (  int    id, double    weight, const DocInfo *    info, const DocumentRep *    dRep )  const [inline, virtual]

Score a given term for a given document. Parameters: id  the term id weight  the weight for this term info  the DocInfo for this document dRep  the DocumentRep for this document Returns: P(t|D) * P(s|t)

virtual void XLingRetMethod::scoreCollection (  const QueryRep &    qry, IndexedRealVector &    results )  [inline, virtual]

Score all documents in the collection. The default implementation provided by this class is to call function scoreDoc, thus to score documents one by one. This is inefficient, so usually a subclass should override this method if a more efficient scoring, e.g., based on inverted index, is possible. Reimplemented from RetrievalMethod.

double XLingRetMethod::scoreDoc (  const QueryRep &    qry, int    docID )  [virtual]

Score a document identified by the id w.r.t. a query rep. Implements RetrievalMethod.

double XLingRetMethod::scoreDocVector (  const XLingQueryModel &    qRep, int    docID, FreqVector &    docVector )  [protected, virtual]

void XLingRetMethod::scoreInvertedIndex (  const QueryRep &    qryRep, IndexedRealVector &    scores, bool    scoreAll = false )  [virtual]

virtual void XLingRetMethod::updateQuery (  QueryRep &    qryRep, const DocIDSet &    relDocs )  [inline, virtual]

update the query -- noop Implements RetrievalMethod.

---

Member Data Documentation

double XLingRetMethod::beta [protected]

bool XLingRetMethod::cacheDocReps [protected]

whether or not to cache document representations

PDict& XLingRetMethod::dictionary [protected]

bool XLingRetMethod::docBasedSourceSmooth [protected]

bool XLingRetMethod::docBasedTargetSmooth [protected]

DocumentRep** XLingRetMethod::docReps [protected]

cache document reps.

int XLingRetMethod::docRepsSize [protected]

number of documents plus 1, the size of the docReps array.

double XLingRetMethod::lambda [protected]

double XLingRetMethod::numSource [protected]

double XLingRetMethod::numTarget [protected]

ScoreAccumulator& XLingRetMethod::scAcc [protected]

const Index& XLingRetMethod::source [protected]

Stemmer* XLingRetMethod::stemmer [protected]

const Stopper* XLingRetMethod::stopper [protected]

ScoreAccumulator* XLingRetMethod::termScores [protected]

---

The documentation for this class was generated from the following files:

• XLingRetMethod.hpp
• XLingRetMethod.cpp

---