Supervised vs Unsupervised vs Reinforcement Learning in Python

In supervised learning, models learn from labeled data to predict outcomes. Unsupervised learning finds patterns in unlabeled data without explicit targets. Reinforcement learning trains agents to make decisions by rewarding good actions, often outside sklearn's scope.

Quick Comparison

Here is a quick table comparing supervised, unsupervised, and reinforcement learning based on key factors.

Aspect	Supervised Learning	Unsupervised Learning	Reinforcement Learning
Data Type	Labeled data (input-output pairs)	Unlabeled data (only inputs)	Environment with feedback signals
Goal	Predict labels or values	Discover hidden patterns or groups	Learn actions to maximize rewards
Common Algorithms	Linear Regression, Random Forest, SVM	K-Means, PCA, DBSCAN	Q-Learning, Policy Gradient (not in sklearn)
Output	Predictions or classifications	Clusters or data representations	Action policies
Use Case Example	Spam detection, price prediction	Customer segmentation, anomaly detection	Game playing, robotics
Library Support in Python	Strong support in sklearn	Strong support in sklearn	Mostly outside sklearn (e.g., stable-baselines3)

Key Differences

Supervised learning requires labeled data, meaning each input has a known output. The model learns to map inputs to outputs, making it ideal for tasks like classification and regression.

Unsupervised learning works with unlabeled data. It tries to find structure or patterns, such as grouping similar data points together (clustering) or reducing data dimensions. It does not predict specific outputs.

Reinforcement learning is different: it involves an agent interacting with an environment, learning from rewards or penalties to make decisions. This approach is not directly supported by sklearn but uses other libraries. It focuses on sequential decision-making rather than static data prediction.

Code Comparison

Example of supervised learning using sklearn to classify iris flowers with a Random Forest classifier.

python

from sklearn.datasets import load_iris
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

# Load data
iris = load_iris()
X_train, X_test, y_train, y_test = train_test_split(iris.data, iris.target, test_size=0.3, random_state=42)

# Train model
model = RandomForestClassifier(random_state=42)
model.fit(X_train, y_train)

# Predict and evaluate
predictions = model.predict(X_test)
accuracy = accuracy_score(y_test, predictions)
print(f"Accuracy: {accuracy:.2f}")

Output

Accuracy: 1.00

Unsupervised Equivalent

Example of unsupervised learning using sklearn to cluster iris data with K-Means.

python

from sklearn.datasets import load_iris
from sklearn.cluster import KMeans

# Load data
iris = load_iris()
X = iris.data

# Cluster data
kmeans = KMeans(n_clusters=3, random_state=42)
kmeans.fit(X)

# Show cluster centers and labels
print("Cluster centers:\n", kmeans.cluster_centers_)
print("Labels:\n", kmeans.labels_[:10])

Output

Cluster centers:
 [[5.006 3.428 1.462 0.246]
 [6.85  3.073 5.742 2.071]
 [5.883 2.74  4.39  1.43 ]]
Labels:
 [0 0 0 0 0 0 0 0 0 0]

When to Use Which

Choose supervised learning when you have labeled data and want to predict or classify new data points accurately.

Choose unsupervised learning when you want to explore data structure, find groups, or reduce dimensions without predefined labels.

Choose reinforcement learning when your problem involves learning a sequence of decisions to maximize rewards, such as in games or robotics, but note sklearn does not support it directly.

Key Takeaways

Supervised learning uses labeled data to predict outcomes and is well supported by sklearn.

Unsupervised learning finds patterns in unlabeled data, useful for clustering and exploration.

Reinforcement learning trains agents via rewards for decision-making, typically outside sklearn.

Use supervised learning for prediction tasks, unsupervised for data discovery, and reinforcement for sequential decision problems.

Sklearn provides strong tools for supervised and unsupervised learning but not for reinforcement learning.