some plots

KDEy/experiments.py

@@ -1,3 +1,5 @@
+import os
+
 import numpy as np
 from sklearn.linear_model import LogisticRegression
 from os.path import join
@@ -5,7 +7,7 @@ import quapy as qp
 from quapy.protocol import UPP
 from quapy.method.aggregative import KDEyML
 
-DEBUG = True
+DEBUG = False
 
 qp.environ["SAMPLE_SIZE"] = 100 if DEBUG else 500
 val_repeats  = 100 if DEBUG else 500
@@ -21,7 +23,7 @@ if DEBUG:
     bandwidth_range = np.linspace(0.01, 0.20, 10)
 
 def datasets():
-    for dataset_name in qp.datasets.UCI_MULTICLASS_DATASETS[:4]:
+    for dataset_name in qp.datasets.UCI_MULTICLASS_DATASETS:
         dataset = qp.datasets.fetch_UCIMulticlassDataset(dataset_name)
         if DEBUG:
             dataset = dataset.reduce(random_state=0)
@@ -40,7 +42,8 @@ def experiment_dataset(dataset):
         param_grid={'bandwidth': bandwidth_range},
         protocol=UPP(train_va, repeats=val_repeats),
         refit=False,
-        n_jobs=-1
+        n_jobs=-1,
+        verbose=True
     ).fit(train_tr)
     chosen_bandwidth = modsel.best_params_['bandwidth']
     modsel_choice = float(chosen_bandwidth)
@@ -83,7 +86,10 @@ def plot_bandwidth(val_choice, test_results):
 
         # Mostrar la gráfica
         plt.grid(True)
-        plt.show()
+        # plt.show()
+        os.makedirs('./plots', exist_ok=True)
+        plt.savefig(f'./plots/{dataset_name}.png')
+
 
 
 for dataset in datasets():