Lesson 3: Unsupervised Learning (1 hour)

Learning Objectives

• Understand what unsupervised learning is
• Recognize the difference between supervised and unsupervised learning
• Understand clustering as a key unsupervised technique
• Experience unsupervised learning through hands-on visualization

Materials Needed

• Internet connection
• Visualization tools or demos
• Examples of clustering (images, data points)
• Student notebooks
• Colored markers/pens for activity

Time Breakdown

• Review supervised learning (5 min)
• Introduction to unsupervised learning (15 min)
• Clustering explanation and examples (20 min)
• Hands-on clustering activity (15 min)
• Wrap-up (5 min)

Activities

1. Review Supervised Learning (5 min)

• What is supervised learning?
• What are classification and regression?
• Key point: We had labeled data (answers provided)

2. Introduction to Unsupervised Learning (15 min)

• Definition: Unsupervised learning finds patterns in data without labeled examples (no answers provided)
• Key Concept: It's like exploring without a guide - you find patterns yourself
• When to use: When you don't know what you're looking for, or when labeling data is too expensive/difficult

Real-World Examples:

• Customer segmentation (grouping customers by behavior)
• Finding similar products
• Anomaly detection (finding unusual patterns)
• Organizing photos by similarity
• News article grouping by topic

Key Terms:

• Unlabeled data: Data without correct answers
• Clustering: Grouping similar items together
• Pattern discovery: Finding hidden structures

3. Clustering Explanation and Examples (20 min)

What is Clustering?

• Grouping similar things together
• Finding natural groups in data
• No labels needed - the algorithm discovers groups

Simple Example: Sorting Mixed Items

• Show pile of mixed objects (pens, paper clips, erasers, etc.)
• Ask: "How would you organize these?"
• Students naturally group by similarity (color, type, size)
• This is clustering! No one told you the groups - you found them

Real-World Clustering Examples:

1. Customer Segmentation

   • E-commerce company has customer data
   • Groups customers by shopping behavior
   • Group 1: Frequent buyers, expensive items
   • Group 2: Occasional buyers, sales items
   • Group 3: New customers, exploring
   • Used for: Targeted marketing

2. Image Organization

   • Photo app groups similar photos
   • Finds faces, places, events
   • No labels needed - discovers patterns

3. News Article Grouping

   • Groups articles by topic similarity
   • Discovers themes without categories
   • Used for: Recommendation, organization

Visual Demonstration:

• Show scatter plot of points
• Ask: "How many groups do you see?"
• Show clustering algorithm finding groups
• Compare: Human grouping vs. algorithm grouping

4. Hands-On Clustering Activity (15 min)

Activity: Group Similar Items

Option 1: Physical Activity

• Provide collection of items (mixed shapes, colors, sizes)
• Students work in groups
• Task: Group items by similarity
• Explain your grouping criteria
• Compare: Did different groups use different criteria?
• Discuss: How many groups? How did you decide?

Option 2: Digital Activity

• Use simple clustering demo (if available)
• Or: Give students a list of 20 items
• Students group them into categories
• Share different grouping strategies
• Discuss: Are there "right" answers? Multiple valid ways?

Activity: Clustering Visualization

• Show online clustering demo (if available)
• Or create simple example:
  • Draw dots on board representing customers
  • X-axis: Age, Y-axis: Spending
  • Students identify natural groups
  • Show how algorithm would cluster

Reflection Questions:

• How is this different from supervised learning?
• What makes items "similar"?
• Are there multiple "right" ways to cluster?
• What would happen with more data?

5. Wrap-Up (5 min)

• Key difference: Supervised has labels, unsupervised doesn't
• Clustering is finding natural groups
• When would you use unsupervised learning?
• Preview: Next lesson - Reinforcement learning (learning from rewards)

Differentiation Strategies

• Younger students: Focus on physical sorting activities, simpler examples, visual clustering
• Older students: Explore more complex clustering algorithms, analyze when to use different methods
• Struggling learners: Provide more guidance, use concrete examples, simpler grouping tasks
• Advanced learners: Research k-means clustering, explore dimensionality reduction, analyze clustering quality metrics

Assessment

• Understanding of unsupervised learning concepts
• Participation in clustering activities
• Quality of observations about patterns
• Ability to distinguish supervised vs. unsupervised

---