There are three functions in the use of PCA and NFC fit,fit_transform,transform distinguish between their respective functions. Pass the test, barely understand their differences, and make some notes here.
1.fit_transform is a blend of fit and transform, which is equivalent to calling fit before calling transform.
The 2.transform function must be called after the Fit function or an error will be
3.fit_transform returns the result of a reduced dimension, and is a column-compressed
The 4.fit function returns an algorithm class, but its member variable, COMPONENTS_, has data and appears to be the result of executing the algorithm, but only for row compression. Transpose the data into the FIT function, and the results obtained using fit_transform are not the same, but very close. It is indicated that fit is also a certain feature extraction, but the operation of the line.
As for the detailed differences, no one is currently found. Just recommend using Fit_transform, to use fit, note that this function is for row compression
Fit,fit_transform and transform functions of each algorithm class in Sklearn