关键词: 文本挖掘, 情绪语调变量, 信用风险, 脆弱期权定价

Abstract: From the perspective of company information disclosure, quantitative data intuitively reflects the company's operating and financial status, while descriptive unstructured text information is an effective supplement to quantitative data. This article excavates the text information of credit default from the company's annual report, constructs the sentiment index of the intonation variable to regulate the default threshold of vulnerable options, and improves the classic Klein's vulnerable option pricing model. The research shows that as the index of intonation variables increases, the price of European call and put options shows a decreasing trend. And the closer the index is to 1, the faster the price of the option decreases. It shows that option prices are more sensitive to negative emotions, which is in line with the actual conditions of the financial market. In addition, the empirical research results show that the Klein model that does not consider sentiment indicators tends to underestimate option prices. The introduction of sentiment indicators into the vulnerable option pricing model can more accurately analyze the impact of a company's financial distress on credit risk, and the results are closer to the actual market.

Key words: text mining, emotional tone variables, credit risk, vulnerable option pricing