Footnote:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments September 6, 2019 erstellt
Description:
Chinese Abstract: 如何能将海量非结构化数据为投资所用，从海量的新闻中挖掘有用的信 息，对于投资决策至关重要。本文研究金融知识图谱的自动构建，把多源数据， 包括结构化数据、半结构化数据以及非结构化数据，整合在一起，将金融市场错 综复杂的关系以知识图谱的方式刻画。在此基础上，进一步研究基于知识图谱的 关联实体挖掘，解决财经新闻与股票实体的对应影响关系识别问题。本文最后设 计实验，模拟真实投资策略进行回测，取得了较高的超额收益，验证了本文算法 的有效性。
English Abstract: How to use massive amounts of unstructured data for investment, and to mine useful information from massive news is critical to investment decisions. This paper studies the automatic construction of financial knowledge graph, and integrates multi-source data, including structured data, semi-structured data and unstructured data, to characterize the intricate relationship of financial markets in the form of knowledge graph. On this basis, the related entity mining based on knowledge graph is further studied to solve the problem of identifying the corresponding influence relationship between financial news and stock entities. At the end of the paper, the experiment is designed to simulate the real investment strategy for backtesting, and the higher excess returns are obtained, which verifies the effectiveness of the proposed algorithm