[FEATURE-REQUEST]

[san-vak]

Hi! thanks for the great library

Description 1- get the explained variance by the components after applying pca 2- just export a subset of columns like virtual pca columns

**Is your feature request related to a problem? 1- to determine how much components to extract via pca 2- to reduce the size of the dataset

regards

[JovanVeljanoski]

Hi,

I hope this example shows you what you want to know:

import vaex
import vaex.ml
df = vaex.datasets.iris()

pca = vaex.ml.PCA(features=['sepal_length', 'sepal_width', 'petal_length', 'petal_width'])
df = pca.fit_transform(df)

print(f'Explained variance: {pca.explained_variance_}')
print(f'Explained variance ratio: {pca.explained_variance_ratio_}')

cols_to_export = ['PCA_0', 'PCA_1']
df[cols_to_export].export_hdf5('iris_pca.hdf5', progress='widget')

[san-vak]

thank you so much