Aggregation of existing models.

With the rapid expansion of non-customized data assets, developing reliable and objective methods for their valuation has become essential. However, current evaluation techniques often face challenges such as incomplete indicator systems and an over-reliance on subjective judgment. To address these issues, this study presents a structured framework comprising 17 key indicators for assessing data asset value. A neural network is employed to calculate indicator weights, which reduces subjectivity and enhances the accuracy of the assessment. Additionally, knowledge graph techniques are used to organize and visualize relationships among the indicators, providing a comprehensive evaluation view. The proposed model combines information entropy and the TOPSIS method to refine asset valuation by integrating indicator weights and performance metrics. To validate the model, it is applied to two datasets: Bitcoin market data from the past seven years and BYD stock data. The Bitcoin dataset demonstrates the model’s capability to capture market trends and assess purchasing potential, while the BYD stock dataset highlights its adaptability across diverse financial assets. The successful application of these cases confirms the model’s effectiveness in supporting data-driven asset management and pricing. This framework provides a systematic methodology for data asset valuation, offering significant theoretical and practical implications for asset pricing and management.