Print

Print


***** To join INSNA, visit http://www.insna.org *****
I guess one dataset that can really be helpful is the one that was made available for KDD Cup 2012, Track 1.

http://www.kddcup2012.org/c/kddcup2012-track1

The prediction task involves predicting whether or not a user will follow an item that has been recommended to the user. Items can be persons, organizations, or groups and will be defined more thoroughly below. 

Thanks and Regards 


On Tue, Jan 21, 2014 at 8:22 PM, Denis Parra <[log in to unmask]> wrote:
***** To join INSNA, visit http://www.insna.org *****
Dear Vasiliki,

I work in recommender systems for some years and is not usual to find a dataset with the actual logs of recommendations provided and the users acceptance of the recommendations.

I am not sure how familiar you are with evaluating recommender systems' accuracy but the usual way to do it (at least using an off-line dataset, different story is with a user study) is to obtain a dataset with users and their preferences over certain items (a ground truth where commonly the preferences are ratings or some form of implicit feedback: playcounts, clicks, etc.) and you proceed like in a usual prediction task in data mining: split your data for train/test (or train/validation/test), for the training part you can build any model you want (a recommendation model: collaborative filtering, content-based, hybrid, most-popular, etc.), then you evaluate your predictions over the test part. You report you results after cross-validation.

Usual datasets are Movielens , jester jokes , netflix prize (you can find some of them in grouplens page http://grouplens.org/datasets/ but check also SNAP datasets http://snap.stanford.edu/data/ )

Again, I don't know you expertise evaluating recommenders, but in case you have not much experience I suggest you to read:

- Evaluating collaborative filtering recommender systems by Herlocker et al. (2004) and
- Evaluating Recommender Systems by Guy Shani and Asela Gunawardana (2009)

if you have time, you can also read my book chapter:

Denis Parra, Shaghayegh Sahebi. Recommender Systems: Sources of Knowledge and Evaluation Metrics Chapter 7 in Advanced Techniques in Web Intelligence-2: Web User Browsing Behaviour and Preference Analysis, Ed. Juan Velasquez et al., Springer-Verlag, 2013 (I can send it to you if you can't access it)

Hope that helps. Cheers,

Denis Parra
Assistant Professor, CS Department
School of Engineering, PUC Chile
 



On Tue, Jan 21, 2014 at 9:51 AM, Vasiliki Pouli <[log in to unmask]> wrote:
*****  To join INSNA, visit http://www.insna.org  *****

Dear all,
I am a member of a research group in the National Technical University of Athens and I am working on recommendation systems. I am interested in evaluating the accuracy of recommendations and I am looking for datasets that provide recommendations (any kind) and also provide information about how many of these recommendations are accepted by the users.
I would like to ask if anyone is aware of such a dataset.

Thank you.
Vasiliki

_____________________________________________________________________
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.

_____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.



--
Hemank Lamba
http://menetworked.wordpress.com/ 
https://sites.google.com/site/hemanklamba/home
_____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.