Scholarlink

<p>The convertible bond market is an interesting part of the national economy. This paper designs the GTLM-C (GAN-Transformer-LSM-Clauses) model, which first combines the generative adversarial network (GAN) and Transformer to generate returns of stock price by their distribution under the real world probability, then uses volatility as a link to generate stock price paths by their distribution under the risk-neutral world probability, and finally combines the Monte Carlo least squares method with the three clauses to price convertible bonds. Comparing with the Black-Scholes (B-S) method, the finite difference method and the Monte Carlo least squares method with fixed volatility, the GTLM-C model in this paper has a lower error, verifying its effectiveness and superiority in pricing convertible bonds.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Pricing Convertible Bonds Based on GAN and Transformer

  • 期刊论文

  • 发表日期：

    2025年6月23日

查看全文
  • Liang Li,
  • Ziqing Hu,
  • Huiting Guo,
  • Li Wang,
  • Zhigang Zhang,
  • Yongli Xu

摘要

The convertible bond market is an interesting part of the national economy. This paper designs the GTLM-C (GAN-Transformer-LSM-Clauses) model, which first combines the generative adversarial network (GAN) and Transformer to generate returns of stock price by their distribution under the real world probability, then uses volatility as a link to generate stock price paths by their distribution under the risk-neutral world probability, and finally combines the Monte Carlo least squares method with the three clauses to price convertible bonds. Comparing with the Black-Scholes (B-S) method, the finite difference method and the Monte Carlo least squares method with fixed volatility, the GTLM-C model in this paper has a lower error, verifying its effectiveness and superiority in pricing convertible bonds.