Predicting SPARQL Query Execution Time and Suggesting SPARQL Queries Based on Query History

Rakebul Hasan 1 Fabien Gandon 1
1 WIMMICS - Web-Instrumented Man-Machine Interactions, Communities and Semantics
CRISAM - Inria Sophia Antipolis - Méditerranée , Laboratoire I3S - SPARKS - Scalable and Pervasive softwARe and Knowledge Systems
Abstract : In this paper first we address the problem of predicting SPARQL query execution time. Accurately predicting query execution time enables effective workload management, query scheduling, and query optimization. We use machine learning techniques to predict SPARQL query execution time. We generate the training dataset from real queries collected from DBPedia 3.8 query logs. As features of a SPARQL query, we use the SPARQL query algebra operators and different basic graph pattern types that we generate by clustering the training SPARQL queries. We achieved high accuracy (coefficient of determination value of 0.84) for predicting query execution time. Second, we address the problem of suggesting similar SPARQL queries based on query history. Users often need assistance to effectively construct and refine Semantic Web queries. To assist users in constructing and refining SPARQL queries, we provide suggestions of similar queries based on query history. Users can use the suggestions to investigate the similar previous queries and their behaviors.
Liste complète des métadonnées

Cited literature [17 references]  Display  Hide  Download

https://hal.inria.fr/hal-00880314
Contributor : Rakebul Hasan <>
Submitted on : Tuesday, November 5, 2013 - 5:30:51 PM
Last modification on : Monday, November 5, 2018 - 3:52:09 PM
Document(s) archivé(s) le : Friday, April 7, 2017 - 10:00:22 PM

File

RR-8392.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00880314, version 1

Collections

Citation

Rakebul Hasan, Fabien Gandon. Predicting SPARQL Query Execution Time and Suggesting SPARQL Queries Based on Query History. [Research Report] RR-8392, Inria. 2013. ⟨hal-00880314⟩

Share

Metrics

Record views

408

Files downloads

856