基于动态Stacked-GBDT算法的数据资源价值评估方法研究

针对现有的数据资源价值评估与定价方法主观性强、定量标准缺乏的问题，提出基于模型堆叠集成GBDT(Stacked-GBDT)算法的数据资源价值评估方法。首先，基于敏感性分析，从数据自身和市场两个维度归纳并建立了数据资源价值评估指标体系；然后，基于GBDT机器学习算法与Stacking集成学习算法，提出了基于StackedGBDT的数据资源价值评估算法，并与Random Forest和XGBoost算法进行对比以验证所提方法的正确性及有效性；最后，应用Stacked-GBDT模型对数据集进行动态定价。结果表明，Stacked-GBDT算法构建的数据资源价值评估模型可为数据价值测算及动态定价提供精确可靠的依据与支撑。

Research on data resource value assessment method based on dynamic Stacked-GBDT ensemble learning

Data value assessment and pricing is a key issue in the marketization and assetization of data elements, which is crucial to promote the flow of data elements, value mining and the development of digital economy. In response to the problems that the existing data resource value assessment and pricing methods are highly subjective and lack of quantitative standards, a data resource value assessment and dynamic pricing method based on Stacked-GBDT algorithm is proposed. Firstly, based on sensitivity analysis, the data resource value evaluation index system is summarized and established from two dimensions: data itself and market factors; then, based on GBDT machine learning algorithm and Stacking integrated learning algorithm, the Stacked-GBDT-based data resource value evaluation algorithm is proposed, and compared with Random Forest and XGBoost. Finally, the Stacked-GBDT model is applied to dynamically price the data set. The results show that the data resource value evaluation model constructed by the Stacked-GBDT algorithm in this paper can provide accurate and reliable basis and support for data value measurement and dynamic pricing.