Machine Learning: An Applied Econometric Approach [pdf]

scholar.harvard.edu

165 points·lainon··9 comments

This article had two components that I appreciate:

1. It gave us a broad overview of the field with data (example, the relative performance on various algorithms on certain problems)

2. It gave us another way of looking at some math we may be familiar with. For example, it looked at econometrics as parameter estimation of a linear model, and ML as the opposite: mapping a function without estimating meaningful or consistent parameters.

3. I know I only said two, but I also enjoyed some of the examples. They are interesting and they quickly summarize the research history.

See one of the authors, Sendhil Mullainathan, speak about a similar subject here - a lot more high level but he's a great speaker.

https://www.youtube.com/watch?v=W7izsrYGhdU&feature=youtu.be...

Seems to be an "intro to ML for economists" type deal. Should discuss some Bayesian ML though, would be of more interest to them it seems

Very nice explanation of why you can't usually use features selected in ML models for inference tasks on pg 10-12(96-98).

Any article with "AI/ML" and "applied" I'm going to read.

Seriously. I love the coverage on AI/ML but so much of it missed practical applications.

Surely that title is backwards. I really doubt there is any ML methods that are inspired or based on econometric approaches, but there is no doubt that there exist econometric problems which can be approached using machine learning.

Just so you know, economic regression modelling (the main application econometrics) has existed since the late 1800's. ML has existed since the 1950's. So yes, economists, econometricians and mathematicians who have worked on applied economics would have inspired many aspects of ML, and vice versa.

Touche.

Reply on news.ycombinator.com