Ranking documents in response to users' information needs is a challenging task, due, in part, to the dynamic nature of users' interests with respect to a query or similar queries. We hypothesize that the interests of a given user could be similar to the interests of the broader community of which she is a part at the given time and propose an innovative method that uses social media to characterize and model the interests of the community and use this dynamic characterization to improve future rankings. By generating community interest language model (CILM) for a given query, we use community interest to compute the ranking score of individual documents retrieved by the query. The CILM is based on a continuously updated set of recent (daily or past few hours) user-oriented text data while smoothed by historical community interest. The user-oriented data can be user blogs or user generated textual data.
[1]
Xiaozhong Liu,et al.
Computational community interest for ranking
,
2009,
CIKM.
[2]
Alistair Moffat,et al.
Improved Retrieval Effectiveness Through Impact Transformation
,
2002,
Australasian Database Conference.
[3]
CHENGXIANG ZHAI,et al.
A study of smoothing methods for language models applied to information retrieval
,
2004,
TOIS.
[4]
Michael I. Jordan,et al.
Latent Dirichlet Allocation
,
2001,
J. Mach. Learn. Res..
[5]
Jaana Kekäläinen,et al.
Cumulated gain-based evaluation of IR techniques
,
2002,
TOIS.