A Fuzzy Logic-Driven Semantic and Binary Tree-Based Indexing Framework for Scalable IoT Data Storage and Retrieval

Wedi'i Gadw mewn:

Manylion Llyfryddiaeth
Cyhoeddwyd yn:	Informatica vol. 49, no. 24 (Jul 2025), p. 85-111
Prif Awdur:	Halimi, Khaled
Awduron Eraill:	Hadjadj, Abelhalim, Kouahla, Zineddine, Farou, Brahim
Cyhoeddwyd:	Slovenian Society Informatika / Slovensko drustvo Informatika
Pynciau:	Accuracy Interoperability Datasets Classification Internet of Things Data exchange Multilayers Ontology Fuzzy logic Real time Semantic web Data retrieval Heterogeneity Efficiency Data management Health care Hypotheses Decision making Indexing Connectivity Performance degradation Data storage Resource Description Framework-RDF Semantics Information retrieval
Mynediad Ar-lein:	Citation/Abstract Full Text Full Text - PDF
Tagiau:	Ychwanegu Tag Dim Tagiau, Byddwch y cyntaf i dagio'r cofnod hwn!

Disgrifiad
Crynodeb:	The rapid growth of Internet of Things (IoT) devices presents significant data management challenges due to heterogeneity, interoperability issues, and massive data volumes, which hinder seamless data exchange and limit the IoT's potential. While the Semantic Internet of Things (SIoT) offers improvements through semantic web technologies, existing approaches often struggle with scalable data storage and efficient retrieval. To address this, the paper proposes a comprehensive, multi-layered architecture for efficient, scalable semantic IoT data handling. The architecture comprises: (1) an Edge Layer that utilizes the SAREF ontology to standardize heterogeneous device data into RDF format; (2) a Fog Layer performing fuzzy logic-based classification for enhanced data organization under uncertainty and binary tree-based indexing for efficient retrieval; and (3) a Cloud Layer for centralized storage. This approach integrates fuzzy logic for improved data categorization, particularly demonstrated through enhanced MEWS classification in healthcare, and a novel binary tree indexing method optimized for RDF file retrieval based on semantic content and fuzzy scores. Three dedicated algorithms govern the classification, indexing, and retrieval phases. Experimental validation using healthcare datasets demonstrates the framework's effectiveness. Specifically, the binary tree indexing reduces average retrieval times by orders of magnitude compared to non-indexed. Furthermore, the complete framework maintains stable and low query execution times (<0.01 s) even with 100,000 RDF files, significantly outperforming traditional RDF triple stores, which exhibit substantial performance degradation at scale. By significantly improving RDF data organization and retrieval efficiency, this work offers a scalable and innovative solution for managing Big IoT data, paving the way for advancements across various sectors.
ISSN:	0350-5596 1854-3871
DOI:	10.31449/inf.v49i24.8039
Ffynhonnell:	Advanced Technologies & Aerospace Database