Say you have a dataset with only a few categorical variables but each categorical variable has a lot of levels. You dummy encode each categorical variable and give every level a binary variable which increases the number of columns/variables in your dataset. Would this dataset be considered a "high dimensional" dataset even though almost all the columns are binary encoded dummy variables with 0's and 1's?
Does dummy coding produce a "high dimensional" dataset?
- Thread starter amd278
- Start date
- Tags dimensions dummy coding
Hi Amd,
From a logical perspective, I would think about a categorical variable as only one dimension.
Multiple linear regression formula with matrices: (least-squares model)
B=((X'X)^(-1))*X'Y
From the calculation perspective, when you calculate linear regression each additional dummy increases the size of the X matrix.
Each dummy variable is another dimension in X.
From a logical perspective, I would think about a categorical variable as only one dimension.
Multiple linear regression formula with matrices: (least-squares model)
B=((X'X)^(-1))*X'Y
From the calculation perspective, when you calculate linear regression each additional dummy increases the size of the X matrix.
Each dummy variable is another dimension in X.
I don't think there is agreement on this issue.
https://stats.stackexchange.com/questions/204881/what-is-high-dimensional-data-in-data-mining
It depends on how many variables you have and your field
https://stats.stackexchange.com/questions/204881/what-is-high-dimensional-data-in-data-mining
It depends on how many variables you have and your field