Skip to main content
Practice

Dataset Structure: Features and Labels


In machine learning, a dataset is typically organized into:

  • Features (X) – The input variables used by the model to make predictions.
    Examples: age, height, number of purchases.
  • Labels (y) – The target variable the model is trying to predict.
    Examples: whether an email is spam, the price of a house.

A model learns the relationship between features and labels in supervised learning.


Loading a Dataset in Scikit-learn​

Scikit-learn provides built-in datasets for practice. One of the most famous is the Iris dataset.

Loading the Iris Dataset
from sklearn.datasets import load_iris

iris = load_iris()

# Features (X) - shape: (samples, features)
X = iris.data
print("Feature shape:", X.shape)
print("First row of features:", X[0])

# Labels (y) - shape: (samples,)
y = iris.target
print("Label shape:", y.shape)
print("First label:", y[0])

Inspecting Feature and Label Names​

Feature and Label Names
print("Feature names:", iris.feature_names)
print("Target names:", iris.target_names)

Why This Matters​

  • Features are the information your model uses to make predictions.
  • Labels define the correct answers during training.
  • Organizing data correctly into X and y is essential for Scikit-learn functions like train_test_split() and .fit().

Key Takeaways​

  • X → input features, 2D array shape (n_samples, n_features).
  • y → target labels, 1D array shape (n_samples,).
  • Proper separation of features and labels is the first step in preparing data for training.

What’s Next?​

In the next lesson, we’ll learn how to split data into training and testing sets to evaluate model performance.

Want to learn more?

Join CodeFriends Plus membership or enroll in a course to start your journey.