【摘要】 采矿权评估中，未来价值的高度不确定性及传统B-S模型对历史数据依赖的局限性，使得评估精度亟待提升。本文以A公司并购B公司X磷矿采矿权为案例，提出基于蒙特卡洛模拟与机器学习的波动率参数重构方法。通过构建GARCH(1,1)-Student’s t模型捕捉价格波动的时变性与厚尾特征，结合LSTM神经网络挖掘非线性波动规律，并利用蒙特卡洛模拟生成多维随机价格路径，动态优化实物期权模型的波动率参数。研究表明，该方法有效突破传统模型的静态假设，将波动率估计误差降低23%，使案例采矿权评估值从52 443万元提升至57 395万元，为高风险矿业权评估提供了融合动态建模与风险量化的科学解决方案。更多还原

【Abstract】 In the evaluation of mining rights, the high uncertainty of future value and the limitations of traditional B-S models in relying on historical data make it urgent to improve the accuracy of the evaluation. This article takes the acquisition of X phosphate mining rights by Company A and Company B as a case study, and proposes a volatility parameter reconstruction method based on Monte Carlo simulation and machine learning. By constructing a GARCH(1,1)-Student’s t model to capture the time-varying and fat tailed characteristics of price fluctuations,combined with LSTM neural network to mine nonlinear volatility patterns, and using Monte Carlo simulation to generate multidimensional random price paths, the volatility parameters of the real option model are dynamically optimized. Research has shown that this method effectively breaks through the static assumptions of traditional models,reduces the estimation error of volatility by 23%, and increases the evaluation value of mining rights from 524.43 million yuan to 573.95 million yuan, providing a scientific solution for high-risk mining rights evaluation that integrates dynamic modeling and risk quantification.更多还原