🔄 StratifiedKFold Explained with Example (Scikit-Learn)

-

Cross-validation is a powerful technique for evaluating machine learning models. But when working with imbalanced datasets, regular KFold may produce test sets where one class is missing.

That’s where StratifiedKFold comes in: it ensures class proportions are preserved across folds. Let’s walk through a small example and analyze a multiple-choice question about possible training sets.

📌 The Code

import numpy as np
from sklearn.model_selection import StratifiedKFold

X = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9])
y = np.array([0, 1, 0, 1, 0, 1, 0, 1, 0])

kf = StratifiedKFold(n_splits=3)

for train, test in kf.split(X, y):
    print(X[train])

🧮 Step 1: Dataset

  • Features (X) = [1, 2, 3, 4, 5, 6, 7, 8, 9]

  • Labels (y) = [0, 1, 0, 1, 0, 1, 0, 1, 0]

So:

  • Class 0 → [1, 3, 5, 7, 9] (5 samples)

  • Class 1 → [2, 4, 6, 8] (4 samples)

Class ratio = 5 : 4

🧩 Step 2: How StratifiedKFold Splits

  • We set n_splits=3.

  • That means:

    • Each test fold will have 9 ÷ 3 = 3 samples.

    • Each training fold will have 9 − 3 = 6 samples.

  • Stratification ensures both classes appear in every fold (ratios preserved as closely as possible).

🔍 Step 3: Analyze Each Option

Let’s check which training sets are possible.

❌ Option [1, 2, 3, 6, 7]

  • Train size = 5 (too small, should be 6).

  • Invalid.

✅ Option [1, 2, 3, 4, 5, 7]

  • Train labels = [0, 1, 0, 1, 0, 0] → 4 zeros, 2 ones

  • Test labels = [6, 8, 9] → [1, 1, 0] → 2 ones, 1 zero

  • Both train and test contain both classes, sizes are correct.

  • Valid.

❌ Option [2, 4, 6, 8]

  • Train labels = [1, 1, 1, 1] → only class 1

  • Test labels = [1, 3, 5, 7, 9] → only class 0

  • Stratification broken.

  • Invalid.

✅ Option [1, 2, 3, 6, 8, 9]

  • Train labels = [0, 1, 0, 1, 1, 0] → 3 zeros, 3 ones

  • Test labels = [4, 5, 7] → [1, 0, 0] → 2 zeros, 1 one

  • Balanced distribution, correct sizes.

  • Valid.

❌ Option [2, 5, 8]

  • Train size = 3 (too small, should be 6).

  • Invalid.

✅ Option [4, 5, 6, 7, 8, 9]

  • Train labels = [1, 0, 1, 0, 1, 0] → 3 zeros, 3 ones

  • Test labels = [1, 2, 3] → [0, 1, 0] → 2 zeros, 1 one

  • Balanced distribution, correct sizes.

  • Valid.

🎯 Final Answer

The valid training sets that can be printed are:

  • [1, 2, 3, 4, 5, 7]

  • [1, 2, 3, 6, 8, 9]

  • [4, 5, 6, 7, 8, 9]

✨ Key Takeaways

  • Training set size = n_samples - (n_samples / n_splits)
    → here, 6 training and 3 test samples each fold.

  • StratifiedKFold preserves class balance, but doesn’t require the exact same ratio, just ensures both classes are represented.

  • Options with wrong sizes or missing a class are automatically invalid.

👉 Would you like me to also run this code and show the exact 3 folds that scikit-learn produces? That way, you’ll see the real train/test splits to match against the options.

Comments

Post a Comment

Popular posts from this blog

Understanding Data Leakage in Machine Learning: Causes, Examples, and Prevention

-
When building machine learning models, achieving high accuracy on your training or validation data is exciting — but sometimes, that success can be misleading. One common trap that causes overly optimistic results is data leakage . In this blog, we will explore: What is data leakage? Why is it harmful? Real-world examples to understand leakage better How to prevent data leakage in your projects What is Data Leakage? Data leakage happens when your machine learning model has access to information during training that it wouldn’t realistically have when making predictions in the real world. This extra information “leaks” from the future or from the test set, causing the model to cheat. When leakage occurs, your model's performance metrics (like accuracy, R², or F1-score ) look great, but this performance won’t generalize to new, unseen data. Why is Data Leakage a Problem? Overfitting to training data : The model learns shortcuts rather than meaningful pattern...
Read more

🌳 Understanding Maximum Leaf Nodes in Decision Trees (Scikit-Learn)

-
When working with decision trees in machine learning , one of the most common questions is: 👉 How many leaf nodes can a decision tree have, given certain constraints? Let’s walk through a real example. 📌 The Question We initialize a decision tree as follows: from sklearn.tree import DecisionTreeClassifier clf = DecisionTreeClassifier( max_depth =4, random_state =42) We’re told: The dataset has 100 samples and 10 features . The tree is a perfectly balanced binary tree . No pruning is applied. No additional stopping criteria (like min_samples_split , min_samples_leaf , or max_leaf_nodes ). 👉 What is the maximum possible number of leaf nodes in the decision tree? 🌲 Key Concepts 1. Depth vs. Levels A binary tree with max_depth = d has at most d splits from the root to a leaf. Depth d = 4 means there can be 4 splits from the root node down to a leaf node. 2. Maximum Leaf Nodes Formula In a perfectly balanced binary tree : Number of leaf n...
Read more

Linear Regression with and without Intercept: Explained Simply

-
If you’re new to machine learning, linear regression is one of the easiest models to understand. It tries to draw a straight line (or a flat plane, if there are more features) that best fits the data. When we use scikit-learn’s LinearRegression , we can decide whether the line should include an intercept (a constant number added at the end of the equation) or not. The Example Code import numpy as np from sklearn.linear_model import LinearRegression X = np.array([[12, 13], [23, 24], [24, 25], [35, 36], [36, 37], [37, 38]]) # Create target values y y = np.dot(X, np.array([4, 5])) + 6 reg1 = LinearRegression(fit_intercept=True).fit(X, y) s1 = reg1.score(X, y) reg2 = LinearRegression(fit_intercept=False).fit(X, y) s2 = reg2.score(X, y) What’s Happening Here? X is our input data (two features per row). y is the output we want to predict. We made it using this formula: y = 4 ∗ x 0 + 5 ∗ x 1 + 6 y = 4 * x_0 + 5 * x_1 + 6 This formula clearly has an intercept = 6 . Case 1: Wi...
Read more