search

cimentadaj / ml_socsci

A work-in-progress of the notes/book 'Machine Learning for Social Science'
https://cimentadaj.github.io/ml_socsci/
6 stars 0 forks source link

Need a good real world example of K-Means #16

Open cimentadaj opened 4 years ago

cimentadaj commented 4 years ago

You were thinking of perhaps using the example from Bram's document that he sent on family policies in overleaf. That is, trying to see if countries cluster on family policies. This can be extended to include Gosta's example on three worlds of welfare state capitalism.

cimentadaj commented 4 years ago

Just taught this to the students and this was the weakest section. We need a better example. In particular, the slides need to be rearranged. Students get lost when explaining the concept of assinging the points randomly, creating a centroid, reallocating, etc..

Perhaps what we need is a real world example to work it out manually. Perhaps you can ask Chris Bail to reproduce his paper using fuzzing clustering: https://www.jstor.org/stable/25472513#metadata_info_tab_contents

Maybe he has some code to reproduce this. Otherwise, we can also just look at some real world applications using the ESS data, just as he used for his work.