Data asset valuation remains challenging due to intangible, context-dependent characteristics that defy traditional methods. This study proposes a novel AI-driven hybrid framework integrating knowledge graph, preference learning, and Support Vector Regression (SVR). A three-tier indicator system with 18 metrics is constructed via knowledge graph technology, and an AI-based preference learning mechanism reduces subjectivity in weight determination. The framework is validated on Bitcoin (3,413 trading days) and stock index (1,699 trading days) datasets against traditional AHP and standalone SVR benchmarks. Results demonstrate superior predictive performance, reducing RMSE by 1.446 and 1.0673, and MAE by 2.4004 and 0.803, respectively. These findings provide practical tools for enterprise data asset management, portfolio optimization, and cross-sector governance applications.
Gong et al. (Fri,) studied this question.