Rotating: The principle of PCA algorithm explained

Find this article on the Internet, personally feel very clear, learn.

Principle explanation of PCA algorithm

The PCA algorithm reduces the correlation between the components, but the disadvantage is that the dimensionality reduction is not conducive to classifying the data.

The first principle of the algorithm: the meaning of orthogonal basis, covariance, the purpose of diagonalization of matrices, these things are the thing that the mathematical theory is playing rotten, need deep understanding, in different ways, to try the real meaning of these treatments, can not only be entangled in the mathematical formula. In general, mathematics is just a tool for language description, so how to express it as a general description.

All in all, there are some axioms supported by the Discipline Foundation.