![]() |
CiteULike | ![]() |
ChaTo's CiteULike | ![]() |
![]() |
|
![]() |
Register | ![]() |
Log in | ![]() |
Why we search: visualizing and predicting user behaviorIn WWW '07: Proceedings of the 16th international conference on World Wide Web (2007), pp. 161-170.
|
Reviews
[Write a review of this article]
Notes for this articleTime series analysis of query logs. 'Explorer' tool.
Comparing curves for different queries. Map inflection points of one curve into the other, to summarize the differences.
Radar: graphical representation of the differences into a 2D space.
Classes of temporal correlation:
1.- Search tends to increase after blogging activity. 2.- Bloggers don't talk about anticipated events. 3.- There is a certain amount of news until the news becomes really big (e.g.: Enron trial). 4.- Simultaneously (don't know)
Find related articles from these CiteULike users
Find related articles with these CiteULike tags
Posting History
BibTeX record
RIS record