dPULearnPlot

class dPULearnPlot

Bases: object

Plotting class for dPULearn (deterministic Positive-Unlabeled Learning) results [Breimann25].

Visualizes dPULearn results in two ways: the reliable negatives it identifies, shown in a compressed Principal Component Analysis (PCA) feature space (pca()), and an evaluation comparing multiple sets of identified negatives (eval()).

Every plotting method returns a (fig, ax) pair (a thin tuple subclass): unpack as fig, ax = .... For backward compatibility, the returned object also forwards attribute access to ax, so legacy ax = ...; ax.set_title(...) keeps working.

Added in version 0.1.2.

Methods

eval(df_eval[, figsize, dict_xlims, legend, ...])	Plot evaluation output of dPULearn comparing multiple sets of identified negatives.
pca(df_pu[, labels, figsize, pc_x, pc_y, ...])	Principal component analysis (PCA) plot for set of identified negatives.

__init__()

See also

• dPULearn: the logic class whose identified negatives this visualizes.