Service Analytics

6.Principal Component Analysis

best vector

To find the best vector u (u = 1) later we will call this vector u1 we can minimize the sum of squared distances or maximize the sum of squared lengths of the projections. Length of the projection:

Diskussion