< Back to previous page

Project

Query Reverse Engineering in the Context of the Semantic Web. (R-10405)

The exponential growth of data in modern society led us into the present Big Data era, which poses formidable challenges for extracting meaningful information from them. Moreover, current database technologies are often difficult to handle by non-expert users, creating a gap between them and the data. Therefore, novel ways for extracting information are being explored in order to close this gap. Query Reverse Engineering (QRE) is one of them. The learning problem for QRE is, given examples of input and output pairs, to derive a compatible syntactical query. This problem has been widely addressed in the literature for SQL queries on relational databases. Unfortunately, the solutions proposed do not take advantage of new semantic features provided by ontologies describing the data. Also, for graph-based databases following the RDF data model, our knowledge is still limited, which is a pity, because these models are used in the context of the Semantic Web, where most modern data are residing. Our main goal is to develop novel algorithms to derive SPARQL queries for RDF databases as well as extend our fundamental knowledge on the QRE learning problem. Thereto, existing approaches for relational databases will be evaluated for their applicability to graph-based databases. Also, we shall also explore QRE in an incremental setting (improve previously derived queries as new information becomes available) as well as in a distributed environment (the typical setting for Big Data).
Date:1 Feb 2020 →  31 Dec 2023
Keywords:DATA MINING, DATABASE THEORY
Disciplines:Data models
Project type:Collaboration project