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

Brocolli · 2024年08月04日

请教下关于过度拟合是哪种方法

* 问题详情,请 查看题干

NO.PZ202206210100000204

问题如下:

In Remington and Montgomery’s discussion with Winfield on resampling, Montgomery’s comment is most likely:

选项:

A.correct. B.incorrect regarding estimation errors. C.incorrect regarding diversification of asset allocations.

解释:

Solution

C is correct. Montgomery’s comment about the criticisms of resampling is incorrect regarding diversification of asset allocations. Risker asset allocations are over-diversified, not under-diversified. The comment is correct with regard to estimation errors because the asset allocations do inherit the estimation errors in the original inputs.

A and B are incorrect. Risker asset allocations are over-diversified, not under-diversified. However, the asset allocations do inherit the estimation errors in the original inputs.

老师我记得讲过一个过度拟合的问题,但是在讲义里没有找到。可以帮我回忆下是在哪个方法里吗?多谢

1 个答案

Lucky_品职助教 · 2024年08月05日

嗨,从没放弃的小努力你好:


同学你好:


在AA科目中,Asset-Only Asset Allocations 中的 Resampled MVO 中有提到关于过度拟合的问题。


resampling这种方式,就是不断反复的从一个数据集中进行抽样,然后通过蒙特卡洛模拟来进行最优化,一直到能够得到一个让人满意的结果。这方方式的缺点就是增加了过度拟合的可能性,而过度拟合就会使风险资产的风险不能完全表现在抽样模拟的结果内,从而会使风险更大的资产过度分散化,也就是对波动性的预估更平滑。因为重复抽样、重复统计的次数太多了,为了充分分散化甚至重复画了很多次有效前沿,但导致的结果反而没有经济学意义。


进一步讲,over-diversified 过度分散化是Resampling的缺点之一,而过度拟合则是造成over-diversified 的原因。这一点我们在讲义里确实没有体现,同学可以记忆一下。



 

----------------------------------------------
虽然现在很辛苦,但努力过的感觉真的很好,加油!

  • 1

    回答
  • 1

    关注
  • 47

    浏览
相关问题

NO.PZ202206210100000204 问题如下 In Remington anMontgomery’s scussion with Winfielon resampling, Montgomery’s comment is most likely: A.correct. B.incorreregarng estimation errors. C.incorreregarng versification of asset allocations. SolutionC is correct. Montgomery’s comment about the criticisms of resampling is incorreregarng versification of asset allocations. Risker asset allocations are over-versifie not unr-versifie The comment is correwith regarto estimation errors because the asset allocations inherit the estimation errors in the origininputs. A anB are incorrect. Risker asset allocations are over-versifie not unr-versifie However, the asset allocations inherit the estimation errors in the origininputs. 如题,over-versifie指对risker asset 的配置变得更少么

2023-12-18 00:46 1 · 回答

NO.PZ202206210100000204 问题如下 In Remington anMontgomery’s scussion with Winfielon resampling, Montgomery’s comment is most likely: A.correct. B.incorreregarng estimation errors. C.incorreregarng versification of asset allocations. SolutionC is correct. Montgomery’s comment about the criticisms of resampling is incorreregarng versification of asset allocations. Risker asset allocations are over-versifie not unr-versifie The comment is correwith regarto estimation errors because the asset allocations inherit the estimation errors in the origininputs. A anB are incorrect. Risker asset allocations are over-versifie not unr-versifie However, the asset allocations inherit the estimation errors in the origininputs. C错哪了

2022-11-05 02:45 1 · 回答