文中基于情景模拟方法，提出了一种参考历史数据的修正模型。通过构成历史情景，在最大似然原则下确定情景集合的关键参数Θ。通过一个利率产品的数值实验，对原模型、修正模型与传统蒙特卡罗模型的在值风险估计值的精度进行比较，得到了理想的结果。

The paper presented an improved model with historical information adjusting on the Scenario Simulation's method. By constructing some history scenarios, the state key parameter was estimated in the whole scenario set by Maximum Likelihood. A numerical example of interest-rate-related product was used to compare the precisions of the VaR estimator among utilizing the initial model, improved model and the BruteForce MC model. The preferable result is achieved.