中文摘要：

VaR（Value at Risk）是商业银行市场风险管理的重要方法。本文提出了一种基于组合权重伪似然函数的VaR半参数估计方法,选取股票市场数据进行实证分析,并依据新资本协议对市场风险内部模型法的要求及其它VaR检验准则检验了模型结果。结果表明,相对常用VaR模型及改进之前的模型,此方法在多种检验标准下都具有明显优势。

英文摘要：

VaR （Value at Risk）is an important method of market risk management in commercial banks. We propose a semiparametric estimation of VaR based on pseudo-likelihood function of combined weights, do empirical analysis using some stock market data, and test the model results according to requirements for market risk internal models approach of the new Basel accord and other VaR test criterions. The results show that this model has obvious advantages under several test criterions relative to common VaR models and models before improvement.