plot_aucplot_script with examples#
An example showing the aucplot function
used by a scikit-learn regressor.
9 # Authors: The scikit-plots developers
10 # SPDX-License-Identifier: BSD-3-Clause
Import scikit-plot
14 import scikitplot.snsx as sp
18 ax = sp.aucplot(
19 x=[0, 1, 1, 0, 1, 1, 0, 1, 1, 0],
20 y=[0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0],
21 )
25 ax = sp.aucplot(
26 x=[0, 1, 1, 0, 1, 1, 0, 1, 1, 0],
27 y=[0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0],
28 kind="pr",
29 )
33 import matplotlib.pyplot as plt
34 import numpy as np; np.random.seed(0) # reproducibility
35 import pandas as pd
36
37 from sklearn.datasets import (
38 load_breast_cancer as data_2_classes,
39 load_iris as data_3_classes,
40 load_digits as data_10_classes,
41 )
42 from sklearn.linear_model import LogisticRegression
43 from sklearn.model_selection import train_test_split
Load the data
48 X, y = data_10_classes(return_X_y=True, as_frame=False)
49 # X, y = data_3_classes(return_X_y=True, as_frame=False)
50 X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=0)
51 np.unique(y)
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
Create an instance of the LogisticRegression
55 model = (
56 LogisticRegression(max_iter=int(1), random_state=0)
57 .fit(X_train, y_train)
58 )
59 # Perform predictions
60 y_val_prob = model.predict_proba(X_val)
61 # Create a DataFrame with predictions
62 df = pd.DataFrame({
63 "y_true": y_val==1, # target class (0,1,2)
64 "y_score": y_val_prob[:, 1], # target class (0,1,2)
65 # "y_true": np.random.normal(0.5, 0.1, 100).round(),
66 # "y_score": np.random.normal(0.5, 0.15, 100),
67 # "hue": np.random.normal(0.5, 0.4, 100).round(),
68 })
72 ax = sp.aucplot(x=df.y_true, y=df.y_score)
76 ax = sp.aucplot(
77 df,
78 x="y_true",
79 y="y_score",
80 kind="pr",
81 label=f"class 1",
82 )
86 for i in range(10):
87 ax = sp.aucplot(
88 x=y_val==i,
89 y=y_val_prob[:, i],
90 label=f"class {i}",
91 )
92
93 # --- Collect unique handles and labels ---
94 handles, labels = ax.get_legend_handles_labels()
95 by_label = dict(zip(labels, handles)) # deduplicate
96
97 # Override legend
98 ax.legend(by_label.values(), by_label.keys(), title="Legend")
102 for i in range(10):
103 ax = sp.aucplot(
104 x=y_val==i,
105 y=y_val_prob[:, i],
106 kind="pr",
107 label=f"class {i}",
108 )
109
110 # # With raw arrays (no DataFrame)
111 # # Works because seaborn normalizes arrays internally
112 # np.random.seed(i) # reproducibility
113 # ax = sp.aucplot(
114 # x=np.random.normal(0.5, 0.1, 100).round(),
115 # y=np.random.normal(0.5, 0.1, 100),
116 # kind="pr",
117 # label=f"{i}",
118 # )
119
120 # --- Collect unique handles and labels ---
121 handles, labels = ax.get_legend_handles_labels()
122 by_label = dict(zip(labels, handles)) # deduplicate
123
124 # Override legend
125 ax.legend(by_label.values(), by_label.keys(), title="Legend")
Total running time of the script: (0 minutes 0.710 seconds)
Related examples
