Abstract

The Semantic Web has addressed the development of technologies for efficient data exchange through formats such as RDF and the use of SPARQL as a standard query language. However, users often face difficulties in obtaining unexpected or erroneous results without a clear understanding of the source of the problem, resulting in wasted time. In this context, a method for validating SPARQL queries, divided into five phases executed before processing in RDF engines, is proposed. This method is based on an ontology and considers both syntactic and semantic validations, generating detailed error reports to prevent problems during execution. Experimental results show the efficiency of the approach: when querying a dataset of 16,005 triples, syntactic validation did not exceed 0.4508 ms and semantic validation did not exceed 6.2343 ms. In contrast, queries without validation presented longer response times, with differences of up to 68.2492 ms in syntactic validation and 144.4288 ms in semantic validation. This approach not only improves query accuracy and efficiency, but also significantly reduces time and errors in information retrieval in semantic environments. © 2025 IEEE.