Some wisdom about supervised machine learning

During the 4th Italian Workshop of Econometrics and Empirical Economics: “Climate and Energy Econometrics”, I had the chance to attend to the presentation of Giovanni Cerulli’s book on supervised machine learning: Fundamentals of Supervised Machine Learning: With Applications in Python, R, and Stata.

I keep with me some wise words during Giovanni Cerulli’s keynote after a very relevant question of Luca Fanelli about the role of theory in machine learning: “there is an optimal model for prediction (i.e., OLS) and an optimal model for causal inference (i.e., IV)… They might be different models. Don’t use IV for predictions”. Thanks, Giovanni, for this food for thought.

Giovanni also mentioned an interesting point about the different software programs used in the book: “Python is for Engineers, R is for Statisticians and Stata is for Social scientists”.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.