Tuesday, July 5, 2011

Best Paper Award at Hypertext 2011

Our paper ’Beyond the Usual Suspects: Context-Aware Revisitation Support’  has won the Engelbart Best Paper Award at the 22nd ACM Conference on Hypertext and Hypermedia.

According to the jury report, the paper deals with a classic, relevant topic with potentials beyond what is stated in the paper. The paper presents an elegant prototype and a careful two-level evaluation of the approaches

The Engelbart Best Paper Award is named after Hypertext Pioneer Douglas Engelbart (born 1925).

See our previous blogpost for a short summary of the paper (or download the pdf).

 (picture taken by Paul De Bra)

Contact us if you would like to use the Web History Repository for your research.

Tuesday, May 24, 2011

Predicting page revisits: A matter of combining evidence

We have some results to share with you. In the past few months, we performed some experiments with the Web History Repository to find out how we can better support Web users in carrying out repetitive tasks. And we will present the results during the upcoming conferences Hypertext 2011 and the ICWE 2011 conference on Web Engineering. In this blog, we explain the results in a nutshell.

To a certain extent, Web users like you are pretty predictable, but when it comes to irregular activities (such as planning your holidays), it is not very useful to suggest your favorite news sites or forums: most likely, you want to return to this nice hotel booking site with good offers. But what was the URL again? And its name? Something with an A....?

We found out that it is still possible to help you to find this hotel booking site, by combining some pieces of evidence:
  • the list of pages that you visit most frequently and the pages that you visited recently are a good starting point, but they only help you in finding your usual sites
  • but if we also take into account your current page or query (for example, 'hotel booking') and the links that you typically follow from this page, our suggestions become much better
  • and our suggestions improve even more if we take the day of week into account (if you are like me, there's nothing better than looking for nice holiday destinations on a rainy Sunday)
How exactly we combined these pieces of evidence, and how good the predictions actually are, can be read in the conference papers.
  • Ricardo Kawase, George Papadakis, Eelco Herder and Wolfgang Nejdl. Beyond the Usual Suspects: Context-Aware Revisitation Support. Proc. Hypertext 2011 (pdf)
  • George Papadakis, Ricardo Kawase, Eelco Herder, Claudia NiederĂ©e A Layered Approach to Revisitation Prediction. Proc. ICWE 2011 Intl. Conf. Web Engineering, 2011 (pdf)
For those who prefer not to read academic literature with nasty formulas, I will share some more findings and observations soon.

Firefox 4 users: you can now contribute as well with the updated Web History Repository Extension.

Wednesday, February 23, 2011

It's getting bigger and bigger: new dataset available

We are happy to announce the third update of the Web History Repository. The new version contains the anonymized usage logs of 356 contributors, with a total of 2.392.116 page visits. Thanks again to all who contributed.

If you want to make use of the dataset, simply go to the download page.

Spread the word and keep the Repository growing.

Friday, February 18, 2011

Short Interim Update

The Web History Repository has been growing steadily in the past few months. Now that we passed the new threshold of 350 contributors, we will clean the data and publish the next update of the dataset soon.

Using the Repository, we further explored recurrent behavior on the Web. Among other things, we found that most revisits are related to irregularly returning tasks - such as online shopping, travel planning and background research. These are things that we do only every once in a while, because we want to or because we need to. And if you add all the time spent on these irregularly returning tasks, this outnumbers by far the time you killed while visiting your favorite news sites, online communities or other portals. Somehow, that's a very reassuring thought.

Stay tuned for the next update and keep the repository growing. Thanks!

Friday, November 26, 2010

New dataset available

We are happy to announce the latest update of the Web History Repository. The new version contains the anonymized usage logs of 210 contributors, with a total of 1.324.041 page visits. Thanks again to all who contributed.

If you want to make use of the dataset, simply go to the download page.

Spread the word and keep the Repository growing.

Thursday, November 18, 2010

Podcast about the Web History Repository

The word is spreading - several researchers already approached us for the Web History Repository (don't be shy, ask for it, if you plan to use it).

The repository also attracted the attention of my colleagues from the Stellar NoE on technology-enhanced learning. Browser history tools are important for personal information management (the way people acquire, organize, maintain, retrieve and use information items). And personal information management on its turn becomes more and more important for learning - which typically involves searching and using various Web resources - including discussion forums and social networking sites.

I had a very pleasant chat with Erik Duval. Apart from the project itself, we talked about the relation between learning and personal information management, strategies to attract volunteers using social networks, privacy issues to take care of, and the datatel initiative.

Here is the link to the Podcast.

Friday, October 29, 2010

As promised

Thanks to your many contributions the Web History Repository has grown substantially - and it is still growing. As promised, we make the repository available to the research and developer communities. We are happy to announce that the first, cleaned, version of the WHR is available now.

If you want to make use of the dataset, simply go to the download page.