IMPROVING NEWS RANKING BY COMMUNITY TWEETS Xin Shuai, Xiaozhong - - PowerPoint PPT Presentation
IMPROVING NEWS RANKING BY COMMUNITY TWEETS Xin Shuai, Xiaozhong Liu, Johan Bollen Sunday, April 15, 2012 Different information needs by the same short query Will Obama sing at the next American Idol? Alice What's Will Obama Obama's
Sunday, April 15, 2012
Alice Bob
Carl
What's Obama's policy towards China?
Will Obama sing at the next American Idol?
Will Obama lower taxes?
Sunday, April 15, 2012
Demographic features Age Profession Gender Location ......
Sunday, April 15, 2012
User interests
Sunday, April 15, 2012
tweets from C1 tweets from C2 C2 interest C1 interest
ranked news documents from search engine updated ranking for C1 updated ranking for C2 C1 and C2 are two different communities! V
V
Sunday, April 15, 2012
Re-Ranking Based on CA tweets 1 2 3
12:00am, 2011-01-30 2011-01-30, CASunday, April 15, 2012
Queries: 50 popular queries were manually identified through Google Insight Tweets: Daily tweets for two weeks were collected from three states: California, New Y
News documents: Top 10 ranked retrieval documents from Google & Y ahoo News were collected every day Turk evaluations:
Sunday, April 15, 2012
total # of Y ahoo news document evaluated on MTurk total # of Google news documents evaluated
# of tweets/day
5320 news documents from 33 queries are evaluated by 105 Turkers
Sunday, April 15, 2012
CA NY TX
Sunday, April 15, 2012
Why our method does not work well for TX Better representation of users interests: named entity, historical effect, ... Query subject difference Other demographic features to form community
Sunday, April 15, 2012