开发者:上海品职教育科技有限公司 隐私政策详情

应用版本:4.2.11(IOS)|3.2.5(安卓)APP下载

PASS · 2023年09月16日

正態分佈 vs. 多元正態分佈

* 问题详情,请 查看题干

NO.PZ202110140100000407

问题如下:

To address Concern 3 when designing Approach 1, Yuen should:

选项:

A.add monthly return observations to the dataset using interpolation.

B.randomly sample from the historical returns with replacement.

C.choose the multivariate normal distribution as the initial functional form.

解释:

B is correct. Random sampling with replacement, also known as bootstrapping, is often used in historical simulations because the number of simulations needed is often larger than the size of the historical dataset. Because Approach 1 is a historical simulation and Concern 3 notes that the number of simulations needed is larger than the size of the historical dataset, bootstrapping should be used.

A is incorrect because this approach would result in creating observations that do not exist in the historical record. Doing so would violate the assumption and procedures of historical simulation.

C is incorrect because choosing the multivariate normal distribution as the initial functional form is typically done in a Monte Carlo simulation (Approach 2), not in a historical simulation (Approach 1). Historical simulation randomly samples from the historical dataset by drawing a number from a uniform distribution so that there is equal probability of being selected. Choice of distribution would not address the concern about the size of the dataset.

老師,對於正態分佈跟多元正態分佈兩者的差異及應用場景能否再說明一些,我感覺目前只能直覺的反應出多元正態分佈就是蒙特卡洛下使用的,但確切為何要使用,以及什麼情況下使用還是挺模模糊糊的

1 个答案

星星_品职助教 · 2023年09月19日

同学你好,

在组合中,只需要掌握:

1)一个变量(如单一资产的return)服从单一的正态分布。

2)多个变量(如多个资产构成的资产组合)服从多元正态分布。

蒙特卡洛模拟需要根据变量的分布来发射随机数进行模拟,所以,在做模拟之前,必须要确定这个变量的分布是什么。例如要模拟一个资产组合return的情况,就要事先确定return的分布。

很多的情况下,都会以多元正态分布作为起点和基准(然后再做后续调整)。但这并不意味着MCS中只能设定正态分布,也可以设定别的分布进行模拟。

  • 1

    回答
  • 0

    关注
  • 362

    浏览
相关问题