Datasets for the Evaluation of
Linked Open Data
Recommender Systems

Sisinf Lab logo

The advent of the Linked Open Data (LOD) initiative gave birth to a variety of open knowledge bases freely accessible on the Web. They provide a valuable source of information that can improve conventional recommender systems, if properly exploited.
The evaluation of recommender systems has been, and still is, a fundamental topic in the field. While, several datasets are available for evaluating the performance of collaborative filtering and content-based techniques, for semantic-aware and LOD based recommender systems those datasets are missing. Specifically, these semantic-aware techniques rely on the availability of semantic descriptions about the items and/or users in the system. Here, we provide mappings between the items in well known datasets to DBpedia resources. This can allows practitioners in the field to evaluate and compare their algorithms with existing approaches.



DBpedia mappings to MovieLens1M dataset (v1.2 - June 19th, 2016)

DBpedia mappings to MovieLens1M dataset (v1.1 - July 21st, 2015)

DBpedia mappings to MovieLens1M dataset (v1.0)

These are the mappings to the movies in the MovieLens1M dataset, dataset published by GroupLeans research group.

An example of mapping is the following one:
858 Godfather, The (1972) http://dbpedia.org/resource/The_Godfather
The first two fields (858 and Godfather, The (1972)) are respective movie id and movie title in the MovieLens1M dataset. The third field is the correspodent DBpedia URI.

If you use these mappings in your scientific work, please cite as
@Article{DOTD16, 
  author = {Tommaso {Di Noia} and Vito Claudio Ostuni and Paolo Tomeo and Eugenio {Di Sciascio}}, 
  title = "SPRank: Semantic Path-based Ranking for Top-N Recommendations using Linked Open Data", 
  journal = "ACM Transactions on Intelligent Systems and Technology (TIST)", 
  year = "2016", 
  note = "to appear", 
  url = "http://sisinflab.poliba.it/sisinflab/publications/2016/DOTD16" 
}
or
 @inproceedings{DiNoia:2012:LOD:2362499.2362501,
 author = {Di Noia, Tommaso and Mirizzi, Roberto and Ostuni, Vito Claudio and Romito, Davide and Zanker, Markus},
 title = {Linked Open Data to Support Content-based Recommender Systems},
 booktitle = {Proceedings of the 8th International Conference on Semantic Systems},
 series = {I-SEMANTICS '12},
 year = {2012},
 isbn = {978-1-4503-1112-0},
 location = {Graz, Austria},
 pages = {1--8},
 numpages = {8},
 url = {http://doi.acm.org/10.1145/2362499.2362501},
 doi = {10.1145/2362499.2362501},
 acmid = {2362501},
 publisher = {ACM},
 address = {New York, NY, USA},
} 



DBpedia mappings to Last.fm hetrec2011 dataset (v1.2 - June 19th, 2016)

DBpedia mappings to Last.fm hetrec2011 dataset (v1.1 - July 21st, 2015)

DBpedia mappings to Last.fm hetrec2011 dataset (v1.0)

These are the mappings to the musical artists in the hetrec2011-lastfm-2k dataset.

Here, another example of mapping:
1934 Adele http://dbpedia.org/resource/Adele
The first two fields (1934 and Adele) are respective artist id and artist name in the Last.fm hetrec2011 dataset. The third field is the correspodent DBpedia URI.

If you use these mappings in your scientific work, please cite as
@Article{DOTD16, 
  author = {Tommaso {Di Noia} and Vito Claudio Ostuni and Paolo Tomeo and Eugenio {Di Sciascio}}, 
  title = "SPRank: Semantic Path-based Ranking for Top-N Recommendations using Linked Open Data", 
  journal = "ACM Transactions on Intelligent Systems and Technology (TIST)", 
  year = "2016", 
  note = "to appear", 
  url = "http://sisinflab.poliba.it/sisinflab/publications/2016/DOTD16" 
}
or
@inproceedings{Ostuni:2013:TRI:2507157.2507172,
 author = {Ostuni, Vito Claudio and Di Noia, Tommaso and Di Sciascio, Eugenio and Mirizzi, Roberto},
 title = {Top-N Recommendations from Implicit Feedback Leveraging Linked Open Data},
 booktitle = {Proceedings of the 7th ACM Conference on Recommender Systems},
 series = {RecSys '13},
 year = {2013},
 isbn = {978-1-4503-2409-0},
 location = {Hong Kong, China},
 pages = {85--92},
 numpages = {8},
 url = {http://doi.acm.org/10.1145/2507157.2507172},
 doi = {10.1145/2507157.2507172},
 acmid = {2507172},
 publisher = {ACM},
 address = {New York, NY, USA},
} 


DBpedia mappings to LibraryThing dataset (v1.2 - June 19th, 2016)

DBpedia mappings to LibraryThing dataset (v1.1 - July 21st, 2015)

DBpedia mappings to LibraryThing dataset (v1.0)

These are the mappings to the books in the LibraryThing dataset.

Here, another example of mapping:
3123767 The Da Vinci Code http://dbpedia.org/resource/The_Da_Vinci_Code
The first two fields (3123767 and The Da Vinci Code) are respective book id and book title in the LibraryThing dataset. The third field is the correspodent DBpedia URI.

If you use these mappings in your scientific work, please cite as
@Article{DOTD16, 
  author = {Tommaso {Di Noia} and Vito Claudio Ostuni and Paolo Tomeo and Eugenio {Di Sciascio}}, 
  title = "SPRank: Semantic Path-based Ranking for Top-N Recommendations using Linked Open Data", 
  journal = "ACM Transactions on Intelligent Systems and Technology (TIST)", 
  year = "2016", 
  note = "to appear", 
  url = "http://sisinflab.poliba.it/sisinflab/publications/2016/DOTD16" 
}
or
 @inproceedings{DiNoia:2012:LOD:2362499.2362501,
 author = {Di Noia, Tommaso and Mirizzi, Roberto and Ostuni, Vito Claudio and Romito, Davide and Zanker, Markus},
 title = {Linked Open Data to Support Content-based Recommender Systems},
 booktitle = {Proceedings of the 8th International Conference on Semantic Systems},
 series = {I-SEMANTICS '12},
 year = {2012},
 isbn = {978-1-4503-1112-0},
 location = {Graz, Austria},
 pages = {1--8},
 numpages = {8},
 url = {http://doi.acm.org/10.1145/2362499.2362501},
 doi = {10.1145/2362499.2362501},
 acmid = {2362501},
 publisher = {ACM},
 address = {New York, NY, USA},
} 
or
@inproceedings{Ostuni:2013:TRI:2507157.2507172,
 author = {Ostuni, Vito Claudio and Di Noia, Tommaso and Di Sciascio, Eugenio and Mirizzi, Roberto},
 title = {Top-N Recommendations from Implicit Feedback Leveraging Linked Open Data},
 booktitle = {Proceedings of the 7th ACM Conference on Recommender Systems},
 series = {RecSys '13},
 year = {2013},
 isbn = {978-1-4503-2409-0},
 location = {Hong Kong, China},
 pages = {85--92},
 numpages = {8},
 url = {http://doi.acm.org/10.1145/2507157.2507172},
 doi = {10.1145/2507157.2507172},
 acmid = {2507172},
 publisher = {ACM},
 address = {New York, NY, USA},
} 


If you are interested in knowing more about how to use this data, you can have a look at these papers:



People

:: hosted by SisInf Lab http://sisinflab.poliba.it ::