Couldn't interpret Categorical variables
in
Linear Algebra and Feature Selection
/
PCA Covariance Matrix in Jupyter – Analysis and Interpretation
How does using Categorical variables justifies the principle of PCA ?
How can we transform categorical variables using the characteristics of standard normal distribution ?
It doesn't make sense to represent the categorical variables which have no position in geometry ?
It could lead to loss of information .
1 answers ( 0 marked as helpful)
Hi Swarntam!
Thanks for reaching out!
In this particular case, categorical variables are part of the data we're working on and we convert them into dummies in order to participate in the process of building principal components. Categorical data is not from a normal distribution since the categories don't have an order.
Hope this helps but feel free to post another question if needed. Thank you.
Best,
Ivan