Just take the xcel sheet from hw 1 you dont need to do HW1 (zipfile) only HW2 should be solved in matelab
Referring to the same data set we used in homework 1 (Eltom et al.,
2016). One of the main issues faced the clustering of this data was to define
the importance of each of the given features. A solution to this issue could be
using PCA to reduce the number of dimensions (features) of the given data.
Questions and Requirements
1.
Apply PCA on the original data set
uploaded with homework 1
2.
Plot the scree plot
3.
Find the number of PCs (dimensions)
that accounts for 85%, 90%, 95%, and 99% of the variation of the data
4.
Apply the K-mean on the PCA result,
use, only, the number of dimensions’ accounts for 90% of the variation of the
data
5.
Compare and contract between the clustering using
K-means alone vs. K-means after PCS.
6.
Plot the PC1, PC2, and PC3 in a 3D
plot, do you notice any clear clustering? Why? OR why not?
