Tutorial 13: Logistic Regression — Breast Cancer

Level: Intermediate · Part of: ML Learning Roadmap

← Back to roadmap

Case Study: Tumor Malignancy Diagnosis

Scenario

Pathologists need a second opinion system that classifies tumors as malignant or benign from cell measurements.

Learning Objectives

• Train LogisticRegression for binary classification
• Interpret coefficients as log-odds
• Use predict_proba for risk scores
• Set classification threshold

Prerequisites

• Tutorial 05
• Python 3.9+, scikit-learn, NumPy, pandas, matplotlib

Dataset

load_breast_cancer() — 30 features, binary target

Hands-On Solution

Copy and run the complete script below:

import numpy as np
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report, roc_auc_score

data = load_breast_cancer()
X, y = data.data, data.target

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42, stratify=y
)

scaler = StandardScaler()
X_train_s = scaler.fit_transform(X_train)
X_test_s = scaler.transform(X_test)

clf = LogisticRegression(max_iter=1000)
clf.fit(X_train_s, y_train)

y_prob = clf.predict_proba(X_test_s)[:, 1]
y_pred = clf.predict(X_test_s)

print(classification_report(y_test, y_pred, target_names=data.target_names))
print(f"ROC-AUC: {roc_auc_score(y_test, y_prob):.4f}")

# Custom threshold: flag if malignant probability > 0.3
y_pred_sensitive = (y_prob > 0.3).astype(int)
print("\nLower threshold (0.3) — higher recall for malignant:")
print(classification_report(y_test, y_pred_sensitive, target_names=data.target_names))

Expected Output

When you run the script, you should see evaluation metrics printed to the console. Some tutorials also save .png plot files in the current directory.

Exercises

1. Find the top 5 features by absolute coefficient magnitude.
2. Try class_weight='balanced'. How do precision/recall shift?
3. Plot predicted probability distribution for malignant vs benign.

Key Takeaways

• Logistic regression outputs calibrated probabilities
• Threshold tuning trades precision vs recall
• Scaling is important for stable coefficient interpretation

Navigation

Part of the ML Learning Roadmap — Hands-On with scikit-learn