主题:【原创】谈谈时间序列的平稳性(1) -- 万里风中虎
Don't quite understand here. Why should we start with something wrong while ADF already gave the right answer? My bottom line: OLS should not be used. Not matter how transparent it is to the audience, it's misleading.
This might be true. But you always have the choice not to bore them with your model details but lead them to your conclusion, as you usually do to industry audience.
It's not a tough call between an overly complicated model and a wrong model.
The results are misleading. The discussion is doomed. Trash in, trash out.
PS. I love most of your posts. This one is an exception.
- 相关回复 上下关系8
压缩 2 层
🙂意见很对,所以我说是一个最简单的预测 2 万里风中虎 字117 2010-02-03 15:18:40
🙂It's not a simple model 5 静默各自想拳经 字414 2010-02-04 00:44:13
🙂这当然是个misspecified的模型 4 万里风中虎 字355 2010-02-04 05:52:35
🙂No offence, but...
🙂赞同老虎的从基础谈起 空白 字58 2010-02-04 09:28:12
🙂兄台是讲随机漫步吗? 1 面壁 字160 2010-02-04 04:41:35
🙂Yes,random walk=随机漫步 3 静默各自想拳经 字843 2010-02-04 08:17:57
🙂我个人山寨地用类似"薛定谔的猫"来解释 面壁 字392 2010-02-04 15:36:54