
Last: DR for compressing data

Here DR for visualize data

DV makes better understanding of the data

Huge graph data that maybe contain 50 parameter

Difficult to visualize Â make 50D graph,

Instead represent each vector, represent countriy in R2.

Makes parameters into one vector. Somehow find the correlation and makes 50D input to 2D input

As example z1 = country size. Generalized for GDP countries

z2 = per person GDP

Example Singapore (bottom right, high per person, but small countries)

USA : per person high, countries: large

Find a correlation that takes 2D graph to represent the data (Example here, to generalize it as finding the economic activity and its correlation between each country and people that reside within it

This is how to make 2D graph representation. It's important to make it 2D to make us better seeing the data as a whole.This makes us understand better about what data trying to give us, and makes appropriate action based on the graph.


PC analysis, doing DR, and how to compress data