## Semantic Search Shortcuts Dataset

This dataset has been used to perform the experiments for the paper When Entities Meet Query Recommender Systems: Semantic Search Shortcuts.

We build a dataset consisting in 130 queries extracted from the http://www.europeana.eu query logs, containing a sample of users’ interactions covering two years (from August 27, 2010 to January, 17, 2012).

We split the dataset in three disjoint sets:

• The 50 queries in the first set are short (composed by only one term);
• The second set contains 50 queries having an average length of 4.2 terms;
• The 30 queries in the third set have an average length of 9.93 terms.

For each query in the three sets, we compute the top-10 recommendations produced by the SS query recommender system (see the paper) for more details on the recommender) and we map them to entities by using a simple interface providing an user-friendly way to associate entities to queries. Three annotators participated to the manual annotation. They manually annotated queries and their related recommendations with one or more Wikipedia entities. An entity is represented by its title and its numerical id, available in a recent English Wikipedia dump, e.g.

  wikiName:Krater
id:1502187


represents the entity described here http://en.wikipedia.org/wiki/Krater.

Queries are available here: the file contains the flat list of the queries in the format:

	<id> <tab> <query>


of course, queries with id in [1,50] belong to the first set, queries with id \in [51, 100] belong to the second set while the others belong to the third set.

The annotated recommendations for the each query are available here. Each line of this file represents the annotated recommendations for a query encoded using json.

Each record contains a field query, which has a field query containing the raw query extracted from the logs and a field entities contains the list of entities associated by the annotators (each entity is represented by its wikiname and its wiki-id). The other field in the record is called suggestions and is made by a list of recommended queries, each one modelled as a map containing a field query with the raw query recommended, and a field entities with the entities associated by the annotators.

### Example


{
"query": {
"query": "amphore",
"entities": [
{
"wikiName": "Amphora",
"id": 51812
}
],
"suggestions": [
{
"query": "amphore   hermes",
"entities": [
{
"wikiName": "Amphora",
"id": 51812
},
{
"wikiName": "Hermes",
"id": 14410
}
]
},
{
"query": "amphore antigone",
"entities": [
{
"wikiName": "Amphora",
"id": 51812
},
{
"wikiName": "Antigone",
"id": 67196
}
]
}
]
}
}



contains two recommendations for the query amphore (associated to the entity Amphora. One is the query amphore hermes (associated to the entities Amphora and Hermes. The other is the query amphore antigone ( associated to the entities Amphora and Antigone ).

@inproceedings{ceccarelli2013entities,
title={When Entities Meet Query Recommender Systems: Semantic Search Shortcuts},
author={Ceccarelli, D. and Gordea, S. and Lucchese, C. and Nardini, F.M. and Perego, R.},
booktitle={SAC '13: 28th Symposium On Applied Computing, Coimbra, Portugal, March 2013},
year={2013}
}


Feel free to contact me via email for more details.

Licensed under the Apache License, Version 2.0 (the ‘License’); you may not use this file except in compliance with the License. You may obtain a copy of the License at