| 朝乐门.数据创造价值:特征、方法与影响因素分析[J].中国图书馆学报,2025,51(3):65~86 |
| 数据创造价值:特征、方法与影响因素分析 |
| Value Creation from Data:An Analysis of Features,Techniques,and Determinants |
| 投稿时间：2024-03-08 |
| DOI： |
| 中文关键词: 数据 数据价值 数据创造价值 数据估值矩阵 |
| 英文关键词: Data Data value Data creation value Data valuation matrix |
| 基金项目: |
|
| 摘要点击次数: 275 |
| 全文下载次数: 160 |
| 中文摘要: |
数据创造价值是当今社会的焦点问题之一。为了深入揭示数据创造价值活动的内在机理,必须从价值维度重新审视数据,尤其需要突破对大数据的V's属性和DIKW模型的固有认识,探究大数据的价值属性与其他属性之间的联系,进而建立从数据至认知的直通道。“数据价值分析理论的第一性原理”的提出明确了数据创造价值的研究前提假定及研究侧重点,而“数据价值分析矩阵”的提出为数据价值的量化计算及推理分析提供了代数工具。在数据价值分析矩阵中,数据价值的估计不仅可以采用基准法、成本法、收益法、风险法、范围法、纯度法,而且也可以运用综合评价方法——数据价值的综合估值方法。与传统价值创造不同,数据创造价值活动具备涌现性、外部性、加/减法性和弱可解释性等新特征。为此,“数据创造价值模型”解释了数据创造价值过程的基本原理,并将数据创造价值的实现方法分为两种:需求驱动型价值创造和数据驱动型价值创造。其中,前者适用于“需求在先、数据在后”的场景,主要采用假设检验及分组对照试验等数据试验方法；后者则适用于“数据在先、需求在后”的场景,主要运用模型训练与模式识别以及处方性分析等数据洞察方法。然而,成功实施数据创造价值,除了有效运用“数据价值分析矩阵”和“数据创造价值模型”等理论与工具,还需考虑模型可解释与结果可信、过程敏捷与质量控制、数据文化与数据治理以及外部环境与内部平台等影响因素。图6。表4。参考文献42。
|
| 英文摘要: |
Creating value through data has become an increasingly central theme in today's society. To truly understand how data can create value,we must look at it from a fresh perspective,challenging traditional views on the vast V's of big data and the DIKW model. This approach involves exploring how the value attributes of big data interplay with its other characteristics,thereby forging a clear path from raw data to actionable insights. The introduction of the “first principles of data value analysis” sharpens our research focus,laying out the foundational assumptions for investigating how data creates value. Simultaneously,the “data value analysis matrix” emerges as a powerful tool for quantifying and reasoning about data's worth. This matrix enables the estimation of data value through various methods,such as benchmarking,cost,revenue,risk,scope,purity,and a holistic assessment technique. The main values derived from data,especially big data,are encapsulated into eight categories:monetization,optimization,telling（narrative）,insight,validation,alert,experience,and decision,collectively referred to by the acronym MOTIVATED. These values vary across different fields,contributing to economic,social,technological,and political spheres.
Data value creation marks a departure from traditional value creation processes,characterized by its emergent nature,economic externalities,the possibility of additive or subtractive effects,and weak interpretability. The “data value creation model” demystifies the principles behind data value creation,distinguishing between approaches that are led by specific demands and those driven by the data itself. In scenarios where needs define the data collection,demand driven methods,such as experimentation and hypothesis testing,are key. On the other hand,when data precedes the need,data driven approaches,including data mining and model training,become instrumental in uncovering insights.
Data value creation goes beyond merely applying models like the “data value analysis matrix” and the “data value creation model”. It also requires attention to model interpretability,outcome reliability,process agility,quality control,the broader context of data culture and governance,and the interaction between external environments and internal platforms. At its core,creating value from data is an automated,machine led process that spans from data collection to the delivery of data driven products or services. This process emphasizes the complementary strengths of humans and machines,advocating for a human centered approach in the loop of data value creation. This means engaging deeply with the data and iteratively refining processes to enhance the value extracted from data,underscoring the critical role of human insight in the entire journey of data value creation. 6 figs. 4 tabs. 42 refs.
|
|
查看全文
查看/发表评论 下载PDF阅读器 |
|
|
|
