search

uchicago-computation-workshop / 2018_spring_conference

0 stars 1 forks source link

Ningyin Xu #22

Open bensoltoff opened 6 years ago

rickecon commented 6 years ago

I like the experiment of testing the quality of IV estimates vs. machine learning modeling. You can limit the number of variables in your statistical learning approach by using a regularization parameter that penalizes the number of variables. Or you could do a bootstrap that randomly chooses variables.

jamesallenevans commented 6 years ago

Effect of internal migration on children - more $ but less guidance Chinese data What's the answer?! You can blurt it out :-)

xiaow2 commented 6 years ago

I love this topic since I can relate to it very much! Thank you for defining the terms in the beginning! Where did you get the dataset?

dpzhang commented 6 years ago

Interesting topic and I would love to see the final paper!

mcs2017 commented 6 years ago

What potential instrumental variables do you have? I am also wondering what machine learning methods can be used to find an IV.

siyiii commented 6 years ago

Hi Ningyin! Thanks for your presenting. I am wondering if the data (1997-2011) you used is kind of old.