Data Asset Valuation Driven by Neural Dynamic Data Valuation

As the volume of enterprise and organizational data assets is increasing at a very fast rate, conventional models of data valuation have trouble in capturing the dynamic value properties of data in various business contexts and have no capacity to flexibly accommodate data use behavior and risk aspects. To overcome the issues mentioned above, this paper suggests a data asset value evaluation approach on the basis of neural dynamic modeling. The approach feeds the temporal features of data assets when invoked, processed and reused across different scenarios into a recurrent neural network to build a multidimensional state vector z_t. An overall value index V_t is created with the help of a dynamic weighting process. At the same time, a risk reduction factor ρ_t is also added to make the value adaptively adjusted to reach a dynamic balance between returns and risks. Experiments are done based on temporal properties of data resources derived using simulated and real business log data. Some of the indicators, including invocation frequency, the number of participating tasks, and processing depth, are normalized to be utilized as model training. The findings indicate that the risk intensity of D003 is only 0.05, with an initial composite value of 0.91 and a slight decrease of 0.86 after depreciation. This means that under high activity and cross scenario reuse conditions, its risk impact is not significant, and therefore it still maintains a high value. The risk intensity of D002 and D004 is 0.25 and 0.20, respectively. After depreciation, their values decrease to 0.47 and 0.55 respectively, and then to 0.35 and 0.44. The addition of a risk reduction mechanism can dynamically reshape the value of data assets and enable the model to more accurately evaluate the contribution, utilization efficiency, and potential risks of data. This information implies that the neural dynamic valuation model can be useful to indicate the dynamic value development traits of the data assets in the various business activities and risks, and become a quantifiable and understandable dynamic value measurement instrument to data asset management, data asset pricing, and data asset allocation.