Understanding Overfitting and Underfitting in Machine Learning (With Examples and How to Avoid Them)

-

When you build a machine learning model, two common problems you might face are overfitting and underfitting. Both can prevent your model from making good predictions on new data. Let’s dive into what these problems are, see simple examples, and learn how to avoid them.

What is Overfitting?

Overfitting happens when a model learns the training data too well, including the noise and random fluctuations. This means it performs extremely well on the training data but poorly on new, unseen data (test data or real-world data).

Example:
Imagine you’re trying to predict the price of houses based on size. If your model tries to memorize every small detail (including weird outliers) in the training data, it might create a very complex curve that fits all points exactly. This curve might look perfect on training data but will fail on new houses because it has learned noise, not just the real pattern.

Visual idea:

  • Training data: scattered points

  • Overfitted model: a very wiggly curve passing through every point

What is Underfitting?

Underfitting happens when a model is too simple to capture the underlying pattern in the data. It performs poorly both on the training data and new data.

Example:
If you use a simple straight line to predict house prices but the actual relationship is more complex (say, quadratic or exponential), your model will miss the real trend. It won’t fit the training data well, and it will also fail to predict new data accurately.

Visual idea:

  • Training data: scattered points

  • Underfitted model: a flat or almost flat line missing the trend

How to Detect Overfitting and Underfitting?

You can check model performance on both training data and validation/test data:

Scenario Training Error Test Error Interpretation
Overfitting Low High Model too complex
Underfitting High High Model too simple
Good Fit (Just Right) Low Low Model generalizes well

How to Avoid Overfitting?

  1. Use More Training Data
    More data can help the model learn the true patterns instead of noise.

  2. Simplify the Model
    Choose simpler algorithms or reduce model complexity (e.g., reduce polynomial degree).

  3. Regularization
    Add a penalty for complexity using techniques like L1 (Lasso) or L2 (Ridge) regularization.

  4. Cross-Validation
    Use k-fold cross-validation to make sure your model generalizes well.

  5. Early Stopping
    When training models like neural networks, stop training when validation error starts increasing.

  6. Pruning (for Decision Trees)
    Remove branches that have little importance to avoid overly complex trees.

How to Avoid Underfitting?

  1. Use More Complex Models
    Try models that can capture more complexity like decision trees, random forests, or neural networks.

  2. Feature Engineering
    Add new features or transformations of features that better represent the underlying data.

  3. Decrease Regularization
    Sometimes too strong regularization makes the model too simple.

  4. Train Longer or Tune Hyperparameters
    Train for more epochs or tune hyperparameters for better performance.

Summary Table

Problem Cause Symptoms Solution
Overfitting Model too complex, memorizes noise Low train error, high test error Regularization, simpler model, more data
Underfitting Model too simple, misses patterns High train error, high test error More complex model, feature engineering

Final Thoughts

Balancing between overfitting and underfitting is key to building effective machine learning models. Always validate your model on unseen data and adjust complexity accordingly. Use techniques like cross-validation, regularization, and feature engineering to help your model generalize well.

If you want, I can also help with example code or specific model tuning tips. Would you like that?

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