skvisualizations.validation module#

Scikit-learn-compatible visualizations for model validation.

@author: David Diaz Vico @license: MIT

skvisualizations.validation.classifier_scatter(X, y, fname, pca_n_components=2, **kwargs)[source]#

Classifier scatter.

Classifier scatter plot.

Parameters:

  • X (array-like, shape (n_samples, features_shape)) – The transformed data.

  • y (numpy array of shape [n_samples]) – Target values.

  • fname (str or file-like object) – https://matplotlib.org/api/_as_gen/matplotlib.pyplot.savefig.html

  • pca_n_components (integer, default=2) – Dimension of the PCA projection of X.

  • **kwargs (optional savefig named args)

Return type:

None.

skvisualizations.validation.keras_history_plot(history, fname, **kwargs)[source]#

Keras history plot.

Train loss plotted for each training epoch.

Parameters:
Return type:

None.

skvisualizations.validation.metaparameter_plot(search, param, fname, score='score', log_scale=True, **kwargs)[source]#

Metaparameter plot.

Train and test metric plotted along a meta-parameter search space.

Parameters:

  • search (search object) – Fitted sklearn search object.

  • param (string) – Name of the meta-parameter.

  • fname (str or file-like object) – https://matplotlib.org/api/_as_gen/matplotlib.pyplot.savefig.html

  • score (string) – Name of the metric

  • log_scale (boolean, default=True) – Wether to use a logarithmic scale.

  • **kwargs (optional savefig named args)

Return type:

None.

skvisualizations.validation.regressor_scatter(X, y, preds, fname, **kwargs)[source]#

Regressor scatter.

Regressor scatter plot.

Parameters:

  • X (array-like, shape (n_samples, features_shape)) – The transformed data.

  • y (numpy array of shape [n_samples]) – Target values.

  • preds (numpy array of shape [n_samples]) – Predicted values.

  • fname (str or file-like object) – https://matplotlib.org/api/_as_gen/matplotlib.pyplot.savefig.html

  • **kwargs (optional savefig named args)

Return type:

None.