Compact query term selection using topically related text.

• Introduction
• Methods
• Comparison

Introduction

This paper illustrate a method called PhRank, selecting terms from retrieved documents by a query.The effects follows:

Methods

• Graph construction Input:query Q = {w1,…wn}
  N = {d0 , ….dk } ,Q+the top k documents retrieved
  It considers three relationships between terms: co-occurrence, stemming and association.
  Here is an example:
  
  
  

• Edge weights p(dk|Q) : the probability of the document in which the stems i and j co-occur given Q cijW2 and cijW10: the counts of stem co-occurrence in windows of size 2 and 10inN λ is set to 0.6
  

• Random Walk The probability hij = lij if vi and vj are connected, and hij = 0 otherwise. πjt be the affinity score associated with vj at time t.
  

• Vertex weights wnfavg is the frequency of a word wn in N.
  

• Term ranking

Comparison

The End!
Reference: K.Tamsin Maxwell, W.Bruce Croft:SIGIR 2013: 583-592.Compact query term selection using topically related text.

Author

Typing Theme

Lorem ipsum dolor sit amet, consectetur adipisicing elit. Tempora non aut eos voluptas debitis unde impedit aliquid ipsa.