Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring Weighted Property Approaches for RDF Graph Similarity Measure

Ngoc Luyen Le, Marie-Hélène Abel, Philippe Gouspillou

TL;DR

This work addresses the problem of RDF graph similarity by introducing weighted properties to reflect context-dependent importance of different predicates. It proposes a hybrid similarity framework that combines quantitative object comparisons via a distance-based term $Sim_1$ and qualitative text-based similarity via $Sim_2$, yielding an overall score $Sim(G_1,G_2)$ normalized by the sum of weights. The approach is instantiated in the vehicle domain, where weighted variants $P1$–$P11$ are evaluated against baselines PJ, PS, and P0; results indicate that weighted-property variants, particularly $P11$, achieve higher similarity scores and better reflect perceived graph similarity. The study highlights benefits in accuracy and context-sensitivity while also acknowledging challenges in weight calibration and scalability, and points to future work on systematic weight assignment and broader domain applications. The work contributes a pragmatic mechanism to tailor RDF graph similarity to application needs, enabling more precise knowledge discovery and semantic search in domains with heterogeneous property importance.

Abstract

Measuring similarity between RDF graphs is essential for various applications, including knowledge discovery, semantic web analysis, and recommender systems. However, traditional similarity measures often treat all properties equally, potentially overlooking the varying importance of different properties in different contexts. Consequently, exploring weighted property approaches for RDF graph similarity measure presents an intriguing avenue for investigation. Therefore, in this paper, we propose a weighted property approach for RDF graph similarity measure to address this limitation. Our approach incorporates the relative importance of properties into the similarity calculation, enabling a more nuanced and context-aware measures of similarity. We evaluate our approach through a comprehensive experimental study on an RDF graph dataset in the vehicle domain. Our results demonstrate that the proposed approach achieves promising accuracy and effectively reflects the perceived similarity between RDF graphs.

Exploring Weighted Property Approaches for RDF Graph Similarity Measure

TL;DR

This work addresses the problem of RDF graph similarity by introducing weighted properties to reflect context-dependent importance of different predicates. It proposes a hybrid similarity framework that combines quantitative object comparisons via a distance-based term and qualitative text-based similarity via , yielding an overall score normalized by the sum of weights. The approach is instantiated in the vehicle domain, where weighted variants – are evaluated against baselines PJ, PS, and P0; results indicate that weighted-property variants, particularly , achieve higher similarity scores and better reflect perceived graph similarity. The study highlights benefits in accuracy and context-sensitivity while also acknowledging challenges in weight calibration and scalability, and points to future work on systematic weight assignment and broader domain applications. The work contributes a pragmatic mechanism to tailor RDF graph similarity to application needs, enabling more precise knowledge discovery and semantic search in domains with heterogeneous property importance.

Abstract

Measuring similarity between RDF graphs is essential for various applications, including knowledge discovery, semantic web analysis, and recommender systems. However, traditional similarity measures often treat all properties equally, potentially overlooking the varying importance of different properties in different contexts. Consequently, exploring weighted property approaches for RDF graph similarity measure presents an intriguing avenue for investigation. Therefore, in this paper, we propose a weighted property approach for RDF graph similarity measure to address this limitation. Our approach incorporates the relative importance of properties into the similarity calculation, enabling a more nuanced and context-aware measures of similarity. We evaluate our approach through a comprehensive experimental study on an RDF graph dataset in the vehicle domain. Our results demonstrate that the proposed approach achieves promising accuracy and effectively reflects the perceived similarity between RDF graphs.
Paper Structure (8 sections, 4 equations, 3 figures)

This paper contains 8 sections, 4 equations, 3 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Our proposition
  4. Formulations
  5. Weighted Properties
  6. Similarity Measure Using Weighted Properties
  7. Experiments
  8. Conclusion and Perspectives

Figures (3)

  • Figure 1: A snapshot of an RDF graph represents information about a vehicle model (vo note for Vehicle Ontology)
  • Figure 2: Histogram representing the distribution of similarity scores across different approaches.
  • Figure 3: Correlation heat map of similarity measures across 1000 used vehicles for different approaches (Similarity scores ranging between 0 and 1 imply 0 as entirely distinct and 1 as complete similarity between vehicle items)

Theorems & Definitions (3)

  • Definition 1
  • Definition 2
  • Definition 3