Knowledge-Driven Hierarchical Concept Clustering for Interpretable Data Analysis

Ning Kang

doi:10.1016/j.knosys.2026.115511

Knowledge-Driven Hierarchical Concept Clustering for Interpretable Data Analysis

2026-02-19

0

PRE

AI

DOI：10.1016/j.knosys.2026.115511

Original

Original request for help

Save

Abstract

En 中文

• A knowledge-driven clustering framework (KD-HCC) is developed by integrating kernel-based similarity learning with concept-oriented knowledge representation. • Hierarchical concept structures are constructed to encode bidirectional relationships between objects and linguistic term attributes, enabling transparent reasoning. • A concept-aligned mechanism is introduced to identify semantically coherent and interpretable cluster levels within the hierarchy. • Experiments on linguistic and medical datasets demonstrate improved cluster compactness, structural consistency, and semantic interpretability. • The proposed framework establishes a unified foundation for explainable and knowledge-based data analysis.

Keywords:

Knowledge-Driven Clustering

Hierarchical Concept Structures

Interpretable Data Analysis

Concept-Oriented Representation

Kernel-Based Similarity Learning

Knowledge-Driven Hierarchical Concept Clustering for Interpretable Data Analysis

Abstract

Journal

Organization