Closed ijyliu closed 3 years ago
@nicomarto did you intend to use just one, or more than one? Did the prof. have a recommendation?
I'm pretty sure we just wanted to use 1.
Cut down on number of components, perhaps remove not well correlated ones. Most GDP-related items... no need for headcount ratio, poverty gap, ODA
Current results with just a few gdp related indicators
Currently I'm using just one.
But we could choose an optimal number using cross validation.