Plot confusion matrix from Cross-Validation with F1

Plot confusion matrix from Cross-Validation, with F1 as subplot.

Import librairies

from museotoolbox.ai import SuperLearner
from museotoolbox.cross_validation import RandomStratifiedKFold
from museotoolbox.charts import PlotConfusionMatrix
from museotoolbox import datasets
from sklearn.ensemble import RandomForestClassifier

Load HistoricalMap dataset

X,y = datasets.load_historical_data(low_res=True,return_X_y=True)
field = 'Class'

Create CV

RSKF = RandomStratifiedKFold(n_splits=2,
                random_state=12,verbose=False)

Initialize Random-Forest

classifier = RandomForestClassifier()

Start learning

SL = SuperLearner(classifier=classifier,param_grid=dict(n_estimators=[10,50]))
SL.fit(X,y,cv=RSKF)

Get kappa from each fold

for stats in SL.get_stats_from_cv(confusion_matrix=False,kappa=True):
    print(stats['kappa'])

Out:

0.925091417957069
0.9164468361024596

Get each confusion matrix from folds

cms = []
for stats in SL.get_stats_from_cv(confusion_matrix=True):
    cms.append(stats['confusion_matrix'])
    print(stats['confusion_matrix'])

Out:

[[934  12   0   1   0]
 [ 29 258   0   2   0]
 [  1   0 283   0   0]
 [  1  13   3  49   0]
 [  0   1   0   0   0]]
[[933  13   0   1   0]
 [ 28 247   0  14   0]
 [  0   0 284   0   0]
 [  0  10   1  55   0]
 [  1   0   0   0   0]]

Plot confusion matrix

import numpy as np
meanCM = np.mean(cms,axis=0).astype(np.int16)
pltCM = PlotConfusionMatrix(meanCM.T) # Translate for Y = prediction and X = truth
pltCM.add_text()
pltCM.add_f1()
pltCM.color_diagonal()