discrimintools.fviz_candisc

discrimintools.fviz_candisc(obj, element='biplot', **kwargs)

Visualize Canonical Discriminant Analysis (CANDISC)

Canonical discriminant analysis is a dimension-reduction technique related to principal component analysis and canonical correlation. fviz_candisc provides plotnine based elegant visualization of CANDISC outputs.

Parameters:

• obj (class) – An object of class CANDISC.

• element (str, default = ‘biplot’) – The element to plot from the output, possible values:

  • ‘ind’ for the individuals graphs

  • ‘var’ for the variables graphs (= Correlation circle)

  • ‘biplot’ for biplot of individuals and variables

  • ‘dist’ for the distance graphs

• **kwargs – further arguments passed to or from other functions.

Returns:

p – A object of class ggplot.

Return type:

class

See also

fviz_candisc_biplot

Visualize Canonical Discriminant Analysis (CANDISC) - Biplot of individuals and variables.

fviz_candisc_ind

Visualize Canonical Discriminant Analysis (CANDISC) - Graph of individuals.

fviz_candisc_var

Visualize Canonical Discriminant Analysis (CANDISC) - Graph of variables.

fviz_dist

Visualize distance between barycenter.

Examples

>>> from discrimintools.datasets import load_wine
>>> from discrimintools import CANDISC, fviz_candisc
>>> D = load_wine("train") # load training data
>>> y, X = D["Quality"], D.drop(columns=["Quality"]) # split into X and y
>>> clf = CANDISC()
>>> clf.fit(X,y)
CANDISC()

Graph of individuals…

>>> p = fviz_candisc(clf, "ind") # graph of individuals
>>> print(p)

Fig. 1 Graph of individuals - CANDISC

Graph of variables…

>>> p = fviz_candisc(clf, "var") # graph of variables
>>> print(p)

Fig. 2 Graph of variables - CANDISC

Biplot of individuals and variables…

>>> p = fviz_candisc(clf, "biplot") # biplot of individuals and variables
>>> print(p)

Distance between class barycenter.

>>> p = fviz_candisc(clf, "dist") # graph of distance
>>> print(p)

Fig. 3 Distance between class barycenter - CANDISC