This paper presents the University of Amsterdam’s participation in TREC 2014. For the Contextual Suggestion Track, we experimented with the use of anchor text representations in the language modeling framework, and base our runs either on full ClueWeb12 or the subset of touristic aggregators (e.g., tripadvisor) provided by the organizers of the track. We also look at the eectiveness of priors (in particular, PageRank) and ways of formulating the query based on the context. Our main nding is that the anchor text representation is eective for retrieving candidate attractions, and performs better than a standard document text index. A linear combination of both anchor and document text leads to further improvement. For the Web Track, we continued our experiment with the fusion of anchor text relative to the text-based baseline run. Our main nding is, again, that the combination of anchor and document text leads to improvement, and we demonstrate how the fusion weight can be used as a handle to tune the amount of risk acceptable for the risk sensitive evaluation.
[1]
Stephen E. Robertson,et al.
Effective site finding using link anchor information
,
2001,
SIGIR '01.
[2]
Craig MacDonald,et al.
Terrier Information Retrieval Platform
,
2005,
ECIR.
[3]
Jaap Kamps,et al.
The importance of anchor text for ad hoc search revisited
,
2010,
SIGIR '10.
[4]
Jaap Kamps,et al.
Web-centric language models
,
2005,
CIKM '05.
[5]
Charles L. A. Clarke,et al.
Evaluating Contextual Suggestion
,
2013,
EVIA@NTCIR.
[6]
CHENGXIANG ZHAI,et al.
A study of smoothing methods for language models applied to information retrieval
,
2004,
TOIS.
[7]
Djoerd Hiemstra,et al.
MapReduce for Information Retrieval Evaluation: "Let's Quickly Test This on 12 TB of Data"
,
2010,
CLEF.
[8]
Kevin S. McCurley,et al.
Analysis of anchor text for web search
,
2003,
SIGIR.
[9]
Olivier Chapelle,et al.
Expected reciprocal rank for graded relevance
,
2009,
CIKM.