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

YI YU · 2024年10月14日

计算器

NO.PZ2023091601000019

问题如下:

A quantitative analyst is constructing a stock selection algorithm that will be employed in making intraday trades and uses the annual returns of two utility stocks, stock A and stock B, to test the model’s capacity to capture dependence between stock returns. The 5 years of annual returns data for each stock used in the test are shown in the following table:

The analyst estimates that the sample means of the returns of stock A (μA) and stock B (μB) are 0.146 and 0.138, respectively. What is the unbiased estimate of the sample covariance of stocks A and B?

选项:

A.

0.003828

B.

0.003892

C.

0.004785

D.

0.004865

解释:

D is correct. Using the formula for the sample covariance estimator but dividing by n-1 for an unbiased estimate, we get

𝜎𝐴𝐵=1/(𝑛−1)Σ(𝑅𝐴,𝑖𝜇𝐴)(𝑅𝐵,𝑖𝜇𝐵)

Which is expanded as

1/4[(0.18−0.146)(0.32−0.138)+(0.13−0.146)(0.22−0.138)+(0.04−0.146)(0.00−0.138)+(0.30−0.146)(0.10−0.138)+(0.08−0.146)(0.05−0.138)]=(14)(0.01946)=0.004865

A is incorrect. This is the result when the two means are switched in the summation formula and the multiplier used is 1/5.

B is incorrect. This is the result when the multiplier used is 1/5 instead of 1/(5-1).

C is incorrect. This is the result when the two means are switched in the summation formula.

老师好,这道题目如果按计算器要怎么计算呢

我得出

Sx=0.100896

Sy=0.127593

r=0.384118

我把他们全部陈起来得到0.004945,请问这样哪里不对呢

1 个答案

pzqa39 · 2024年10月14日

嗨,努力学习的PZer你好:


题目要求是计算无偏估计,答案里面给的这种方法就是无偏估计,是基于原始的数据并且用n-1

但是,当同学用自己的这种方法来计算协方差时,实际上假设是用样本标准差的计算方式(基于样本偏差),相关系数 r 的计算假设数据已经标准化并按样本均值调整,这与直接根据原始数据计算的无偏协方差略有不同,样本量较小时,无偏估计和样本估计之间的差异会更加明显。因此,直接使用相关系数公式的方法不一定能准确反映无偏协方差,虽然计算出来答案很接近,但是会有差异。


答案这种方法按计算器的话没有什么办法能快速计算,只能一点一点算,每算一块就在草稿纸上记一下这样。

----------------------------------------------
加油吧,让我们一起遇见更好的自己!