elphick.sklearn_viz.features.principal_components.PrincipalComponents

class elphick.sklearn_viz.features.principal_components.PrincipalComponents(x, color=None)[source]

__init__(x, color=None)[source]

Parameters:

x (DataFrame) – X values to transform and plot.
color (Optional[Series]) – the optional series by which to color the markers

Methods

`__init__`(x[, color])	type x: `DataFrame`
`add_loading_vectors`(fig, loadings)	rtype: `Figure`
`plot_explained_variance`([standardised, title])	Plot the cumulative explained variance by principal component.
`plot_loading_vectors`([standardised, ...])	plot the loading vectors.
`plot_principal_components`([plot_3d, ...])	Create the pca plot
`plot_scatter_matrix`([original_features, ...])	Plot a scatter matrix

Attributes

data

plot_explained_variance(standardised=False, title=None)[source]

Plot the cumulative explained variance by principal component.

Parameters:

standardised (bool) – If True, plot the standardised PCA, where vectors are transformed to zero mean and unit variance.
title (Optional[str]) – Optional plot title

Return type:

Figure

Returns:

plot_loading_vectors(standardised=False, by_color=False, title=None)[source]

plot the loading vectors.

Parameters:

standardised (bool) – If True, plot the standardised PCA, where vectors are transformed to zero mean and unit variance. by_color: If True, plot the loading vectors by color group.
title (Optional[str]) – Optional plot title

Return type:

Figure

Returns:

a plotly GraphObjects.Figure

plot_principal_components(plot_3d=False, loading_vectors=True, standardised=False, title=None)[source]

Create the pca plot

Parameters:

plot_3d (bool) – If True plot the top 3 principal components in 3D, otherwise the top 2 in 2D.
loading_vectors (bool) – If True and plot_type is ‘2D’|’3D’ loading vectors will be displayed.
standardised (bool) – If True, plot the standardised PCA, where vectors are transformed to zero mean and unit variance.
title (Optional[str]) – Optional plot title

Loading vectors are implemented manually rather than with annotations (lines with arrows),: the problem is described well here: https://community.plotly.com/t/set-pca-loadings-aka-arrows-in-a-3d-scatter-plot/72905

plot_scatter_matrix(original_features=False, standardised=False, title=None)[source]

Plot a scatter matrix

Parameters:

original_features (bool) – If True, plot the original features, otherwise plot the principal components.
standardised (bool) – If True, plot the standardised PCA, where vectors are transformed to zero mean and unit variance.
title (Optional[str]) – Optional plot title

Return type:

Figure

Returns: