Oil market risk factor identification based on text mining technology

In recent years, the occurrence and scope of oil market risks are rising. To identify the possible risks of the oil market fully and effectively, we propose a method with which to extract the risk factors from network news based on an LDA model (a text-mining technology). The method is as follows: firstly, using web crawler technology to obtain 18,000 news items, we turn them into structured data, the LDA topic model is then used to extract the optimal topics from the news, finally, the topic and risk factors will correspond in some way. After comparing and integrating the rationality of risk factors, 28 risk factors were identified. The results show that various risk factors are centred around geopolitics, war conflicts, environmental protection, OPEC policies, and market supply and demand. Among them, political conflicts, economic sanctions, and warfare involving oil-producing regions are the most prominent factors, specifically as they relate to Latin America economic sanctions, and the Syrian war. The relationships between these factors is analysed to assess oil market risk factors structure, and analyse a specific part of the oil market supply chain. Compared with traditional expert consultation and brainstorming methods, this method shows the advantages of being more comprehensive, detailed and easy to operate.