BEST of KAUST at TREC-2011: Building effective search in twitter

Jinling Jiang; Lailatul Hidayah; Tamer Elsayed; Hany Ramadan

BEST of KAUST at TREC-2011: Building effective search in twitter

Jinling Jiang^*, Lailatul Hidayah, Tamer Elsayed, Hany Ramadan

^*Corresponding author for this work

Research output: Contribution to conference › Paper › peer-review

Abstract

In our first-ever appearance at TREC, we explore initial ideas on building an effective search tool over tweet stream as a participation in this year's microblog track. Among those ideas are tweet expansion with short representation of terms that frequently co-occur with hashtags and URLs, and re-ranking based on statistics that estimate user popularity (using replies and mentions), tweet popularity, URL popu- larity and user topic authority (using simple user profiles). Initial results show that re-ranking improves the effective- ness while expansion sometimes harms it. Overall, the sys- tem built for the task is indeed a great resource for further extensions and experiments.

Original language	English (US)
State	Published - 2011
Event	20th Text REtrieval Conference, TREC 2011 - Gaithersburg, MD, United States Duration: Nov 15 2011 → Nov 18 2011

Conference

Conference	20th Text REtrieval Conference, TREC 2011
Country/Territory	United States
City	Gaithersburg, MD
Period	11/15/11 → 11/18/11

ASJC Scopus subject areas

General Engineering

Cite this

@conference{c4798dc3279446faac2dcb89ed6f58db,

title = "BEST of KAUST at TREC-2011: Building effective search in twitter",

abstract = "In our first-ever appearance at TREC, we explore initial ideas on building an effective search tool over tweet stream as a participation in this year's microblog track. Among those ideas are tweet expansion with short representation of terms that frequently co-occur with hashtags and URLs, and re-ranking based on statistics that estimate user popularity (using replies and mentions), tweet popularity, URL popu- larity and user topic authority (using simple user profiles). Initial results show that re-ranking improves the effective- ness while expansion sometimes harms it. Overall, the sys- tem built for the task is indeed a great resource for further extensions and experiments.",

author = "Jinling Jiang and Lailatul Hidayah and Tamer Elsayed and Hany Ramadan",

year = "2011",

language = "English (US)",

note = "20th Text REtrieval Conference, TREC 2011 ; Conference date: 15-11-2011 Through 18-11-2011",

}

TY - CONF

T1 - BEST of KAUST at TREC-2011

T2 - 20th Text REtrieval Conference, TREC 2011

AU - Jiang, Jinling

AU - Hidayah, Lailatul

AU - Elsayed, Tamer

AU - Ramadan, Hany

PY - 2011

Y1 - 2011

N2 - In our first-ever appearance at TREC, we explore initial ideas on building an effective search tool over tweet stream as a participation in this year's microblog track. Among those ideas are tweet expansion with short representation of terms that frequently co-occur with hashtags and URLs, and re-ranking based on statistics that estimate user popularity (using replies and mentions), tweet popularity, URL popu- larity and user topic authority (using simple user profiles). Initial results show that re-ranking improves the effective- ness while expansion sometimes harms it. Overall, the sys- tem built for the task is indeed a great resource for further extensions and experiments.

AB - In our first-ever appearance at TREC, we explore initial ideas on building an effective search tool over tweet stream as a participation in this year's microblog track. Among those ideas are tweet expansion with short representation of terms that frequently co-occur with hashtags and URLs, and re-ranking based on statistics that estimate user popularity (using replies and mentions), tweet popularity, URL popu- larity and user topic authority (using simple user profiles). Initial results show that re-ranking improves the effective- ness while expansion sometimes harms it. Overall, the sys- tem built for the task is indeed a great resource for further extensions and experiments.

UR - http://www.scopus.com/inward/record.url?scp=84873549253&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:84873549253

Y2 - 15 November 2011 through 18 November 2011

ER -

BEST of KAUST at TREC-2011: Building effective search in twitter

Abstract

Conference

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this