【摘要】 金融风险分析方法的局限性,如数据来源单一性、数据类型简单化和研究角度片面性等,归根结底是金融大数据组织层面的问题。传统金融大数据的扁平化组织忽略了数据中丰富的知识关联。本文从知识组织的角度出发,对金融知识表示方法进行了探究,将知识表示过程分为知识表示模式层、知识实例层和知识挖掘层。本文首先分析了金融大数据的四种典型关联模式,即分类关联、时空关联、统计关联和事件关联,并从实现方式的差异角度,将各类关联模式归类于静态本体、动态本体和社会本体之中;针对每一类本体,本文提出了相应的实现方案,如复用现有的FIBO本体;最后,基于金融风险识别应用案例进行了详细说明。更多还原

【Abstract】 The limitations of financial risk analysis, such as homogeneous data source, simplistic data types, and one-sided research perspective, are attributed to insufficient representation of big data in finance; the traditional flat data organization ignores the rich knowledge connection among financial data. This paper explores the knowledge representation model of big data on financials from a knowledge organization perspective. It classifies the process of knowledge representation into three layers: knowledge representation mode, knowledge instance, and knowledge mining. First, the paper summarizes the typical connection patterns that exist in the financial domain, such as classification connection, spatial-temporal connec‐tion, statistical connection, and event connection, and categorizes these into static ontology, dynamic ontology, and social ontology according to the differential realization. Then, following the implementation of each ontology, it proposes related implementation schemes as well as the reuse of the existing ontology of Financial Industry Business Ontology(FIBO).Lastly, it demonstrates the representation process using the case of financial risk identification.更多还原