PCA Explained Step by Step

PCA Explained Step by Step

All data points lie in a sub-manifold of the 3 dimentional space. If somehow we find this sub-manifold which can be “unfolded” to a lower dimensional space (2D space in this case), then we can reduce the dimensiontality of the data without losing much information.

Principle

Figure 1. The goal of PCA is reducing the dimentionality of data without losing much information.

Read more