Dear Traian Rebedea

There are many datasets for person name disambiguation

http://www.d.umn.edu/~tpederse/namedata.html

http://www.cs.cmu.edu/~einat/datasets.html

Good Luck

Thanks, Modammed! 

I am looking for a dataset that contains web pages, maybe the results of a query for an ambiguous name, that are manually grouped for each different person. The grouping can be done using Mechanical Turk or some other solution. To conclude, I am not interested to disambiguate the appearance of a name in a given text (like email, news article, etc.), but rather to assign a web page to a specific person.

In the links you have provided, which are really interesting, there is only a small dataset of this kind (http://www.d.umn.edu/~tpederse/Data/ak-web-data-rev.zip). Others have been developed for the WePS workshops, but they are not online anymore. However, researchers have been using them and I thought that maybe someone still has them on their computers.

Please check  you may get data set from the following link:

https://sites.google.com/site/munaga71/Home/datasetlinks

Dear Traian,

I know this is a very late reply to your original message. I stumbled upon this thread by chance and realized that since I finished working on WePS related tasks the evaluation data sets have become harder find online.

The data is still available, although for some reason the links don't show up easily when searching on Google.

I have recently started consolidating the data sets and their documentation at

https://github.com/jotaefea/weps

For the record, the original links to the evaluation data for the three Web People Search evaluation campaigns:

• http://nlp.uned.es/weps/weps-1/weps1-data
• http://nlp.uned.es/weps/weps-2/weps2-data
• http://nlp.uned.es/weps/weps-3/data

Thanks a lot, Javier! I thought that probably the data is still available, but not indexed for some reasons. Meanwhile, I have made some big dataset from big websites like about.me, however they take a different stance than WEPS. However, if anyone is interested, I could make them public.

I suggest you search such data through kaggle, e.g., https://www.kaggle.com/c/yandex-personalized-web-search-challenge