Calculate AUC in Python (With Example)

Renesh Bedre    2 minute read

Area Under the Receiver Operating Characteristic Curve (AUC) is a widely used numerical metric for evaluating and comparing the performance of binary classification models such as logistic regression.

The AUC is the entire area under this Receiver Operating Characteristic (ROC) curve and summarises the performance of the model as a single numerical value.

AUC ranges from 0 and 1. The perfect model will have an AUC of 1 (perfectly distinguishing between the two classes). The random model (equal chances of prediction) will have an AUC of 0.5.

AUC is commonly used for comparing and selecting the better model. The higher the AUC, the better the model.

The advantage of AUC is that it is scale-invariant (independent of the absolute scale of predicted probabilities) and classification-threshold-invariant (considers all possible threshold values).

In Python, the AUC can be calculated using the auc() function from the scikit-learn package.

The following step-by-step example explains how to calculate the AUC in R for the logistic regression

Getting the dataset

Fit the logistic regression model using the sample breast cancer dataset.

This sample breast cancer dataset includes four features (predictors) and outcome [patient is healthy (0) or cancerous (1)].

# import package
import pandas as pd

# load dataset
df = pd.read_csv("https://reneshbedre.github.io/assets/posts/logit/breast_cancer_sample_2.csv")

# view first few rows
# Classification is the outcome with two levels with cancer (1) or healthy (0) patients
df.head(2)
  Age        BMI  Insulin  Leptin  Classification
0   48  23.500000    2.707  8.8071               0
1   83  20.690495    3.115  8.8438               0

Train-Test split

Split the dataset into train and test datasets. We will use the train_test_split() function from the sklearn package to split 75% as training and 25% as test datasets.

The training dataset will be used for training the model and the test dataset will be used for prediction.

# import package
from sklearn.model_selection import train_test_split

# split into training and testing
df_train, df_test = train_test_split(df, random_state = 0)

Fit the logistic regression model

Fit the logistic regression model using training dataset,

# import package
from sklearn.linear_model import LogisticRegression

# get X and y
X_train = df_train[["Age", "BMI", "Insulin", "Leptin"]]
y_train = df_train["Classification"]

# fit the model
fit = LogisticRegression(random_state = 0).fit(X_train, y_train)

Perform prediction

Predict the outcome of the test dataset using fitted model,

# perform prediction
# # get X and y
X_test = df_test[["Age", "BMI", "Insulin", "Leptin"]]
y_test = df_test["Classification"]

# calculate predicted probabilities
pred_probs = fit.predict_proba(X_test)[:, 1]

Calculate AUC

Calculate the AUC using the roc_auc_score() function from the sklearn package. The roc_auc_score() takes truth and predicted values and returns the AUC.

# import packages 
from sklearn.metrics import roc_auc_score

# calculate AUC
roc_auc_score(y_true = y_test, y_score = pred_probs)
# output
0.6078

The AUC of the fitted model is 0.6078.

The closer the AUC to 1, the better the model. AUC of 0.6078 implies that the fitted model has poor discrimination and may not perform well in predicting whether the patient is healthy or cancerous.

Enhance your skills with courses on machine learning




This work is licensed under a Creative Commons Attribution 4.0 International License

Some of the links on this page may be affiliate links, which means we may get an affiliate commission on a valid purchase. The retailer will pay the commission at no additional cost to you.

You may also enjoy