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

我是一条鱼 · 2020年06月28日

问一道题:NO.PZ2015120204000027 [ CFA II ]

问题如下:

Epstein asks Litvenko whether it is possible to identify companies that are likely tooutperform the market in the next period rather than just trying to predict the returnfor a company.The most appropriate way to address the outperformance issue discussed byEpstein and Litvenko is to:

选项:

A.

use a probit model.

B.

add a dummy variable to the regression.

C.

replace one or more of the independent variables with its logarithmictransformation.

解释:

A is correct. A probit model is appropriate for a binary decision, such as whether a securityis likely to outperform the market or not.

B is incorrect. A dummy variable is a qualitative independent variable. A model toclassify outcomes as binary requires a qualitative dependent variable.

C is incorrect. A logarithmic transformation of an independent variable is appropriatewhen its relationship with the dependent variable is non-linear, but it will not changethe form of the dependent variable to binary (outperforms or does not).


请问为什么不能选b呢
2 个答案

星星_品职助教 · 2021年07月28日

@Phoebe

这个问题描述的是针对Y变量,所以首先要定义一个新的Y变量。add Y以后才能考虑如何选择X的问题。

B选项add X是只add X(不考虑Y)的意思,这没法解决Y的问题。

星星_品职助教 · 2020年06月29日

同学你好,

这道题目要求的是whether it is possible to identify companies that are likely to “outperform” the market,也就是只有两种可能,要么这家公司是outperform,要么是not outperform,所以在这里是最终要求的Y变量为二值变量(dummy),而不是X值。在regression中添加二值的X变量也没法解决Y是二值的问题。所以不能选B。注意这里的“add a dummy variable to the regression”只能是add X variable,因为Y变量只有一个,没法再“add”

对于Y是二值变量的问题需要用的是probit/logit model,即选A

Phoebe · 2021年07月28日

但是Y这个outperform的结果可以由不同的X来产生呢,多加一个X那么Y outperform的概率或结果会更准,所以可以add X吧