esem.utils.leave_one_out

esem.utils.leave_one_out(Xdata, Ydata, model='RandomForest', **model_kwargs)

Function to perform LeaveOneOut cross-validation with different models.

Parameters

  • Xdata (array-like of shape (n_samples, n_features)) – Parameter values

  • Ydata (array-like of shape (n_samples,)) – Target values.

  • model ({'RandomForest', 'GaussianProcess', 'NeuralNet'}, default 'RandomForest')

  • model_kwargs (dict) – More arguments to pass to the model.

Returns

output (list of n_samples (truth, prediction, variance) tuples) – which can then be passed to validation_plot()