Information Technology Reference
In-Depth Information
Chapter 6
News Recommendation in Real-Time
Benjamin Kille, Andreas Lommatzsch and Torben Brodt
Abstract
Recommender systems support users facing information overload
situations. Typically, such situations arise as users have to choose between an
immense number of alternatives. Examples include deciding what songs to listen
to, what movies to watch, and what news article to read. In this chapter, we outline
the case of suggesting news articles. This task entails a number of challenges. First,
news collections do not remain relevant unlike movies or songs. Users continue to
request novel contents. Second, users avoid creating consistent profiles thus reject
login procedures. Third, requests arrive in enormous streams. Having short consump-
tion times, users quickly request the next article to read. Handling these challenges
requires adaptations to existing recommendation strategies as well as developing
novel ones.
Coffee Time
Suzanne shiveredwhile looking out of the window. It was one of these coldDecember
afternoons where you just want to stay at home, enjoy a cup of hot coffee, and relax
next to the fireplace in the living room. “I hope Laura and Linda will make it on
time today” she thought, a little worried about the safety of her friends. It was not
the first time for them that they'd miss their little get-together—or “gossip club,”
as her husband Steve used to call it. She always complained when he said that, but
actually, she secretly had to admit that he wasn't too far from the truth in the analysis
of her circle of friends. They really were gossip! Especially Laura seemed to know
everything about everyone in the neighborhood and was more than motivated to share
Search WWH ::
Custom Search