Knowledge Graph-Based Integration of Conversational Advisors and Faceted Filtering.

This article focuses on a novel knowledge graph-based approach to integrate faceted filtering and conversational product advisors in e-commerce search interfaces. The proposed system models relationships between advisor questions and answers, product features, and filter values within a unified graph structure, enabling mutual influence between filtering and advising components and providing transparent explanations of recommendations. Two empirical user studies with a total of 448 participants demonstrated that the integrated system with visual explanations improves user trust, acceptance, and understanding of relevant product features compared to baseline systems where components operate separately. Interaction analyses revealed that users primarily followed advisor recommendations even when applying manual filters, and that presenting explanations and visualizations is crucial for realizing the benefits of integration. The approach highlights the potential for enhancing multi-level decision support in online product search while acknowledging challenges such as the manual effort required to build the knowledge graph and the need for further evaluation in real-world settings.