Benefits of using PCA for dimensionality reduction

Using PCA to reduce the dimension of high-dimensional data, there are a few features:

(1) data from high-dimensional to low-dimensional, because of the variance, similar features will be merged, so the data will be reduced, the number of features will be reduced, which helps to prevent the occurrence of overfitting phenomenon. But PCA is not a good way to prevent overfitting, it is better to regularization the data when preventing overfitting.

(2) using the method of dimensionality reduction, the speed of operation of the algorithm is accelerated;

(3) Reduce the memory space used to store data;

(4) in the process of mapping from X (i) to Z (i), the training data is reduced to dimension, and then the test data or validation data is reduced.

Benefits of using PCA for dimensionality reduction