Next ❯

Introduction to Unsupervised Learning

  • Unsupervised learning is a machine learning approach where algorithms analyze unlabeled data to discover hidden patterns, clusters, and relationships.

  • What is Unsupervised Learning?

    • Learns hidden patterns in data without explicit output labels

    • Goal: Discover structure, groupings, or relationships in data

    Simple definition:

    “Let the machine find patterns in data on its own.”


    Difference Between Supervised & Unsupervised Learning

    Feature

    Supervised Learning

    Unsupervised Learning

    Output

    Labeled (known)

    Unlabeled (unknown)

    Goal

    Predict output

    Find patterns/structure

    Examples

    Predict house price, spam detection

    Customer segmentation, market basket analysis

    Approach

    Learn mapping: Input → Output

    Learn hidden structure in Input data


    Labeled vs Unlabeled Data

    • Labeled Data: Features + Target
       Example:

    Study Hours

    Result

    5

    Pass

    2

    Fail

    • Unlabeled Data: Features only, no target
       Example:

    Study Hours

    Attendance

    5

    85

    2

    60

    • Unsupervised Learning uses only unlabeled data.


    Types of Unsupervised Learning

    1. Clustering

    2. Association

    3. Dimensionality Reduction

    4.1 Clustering

    • Groups similar data points into clusters

    • Each cluster contains similar items

    Common Algorithms:

    • K-Means

    • Hierarchical Clustering

    • DBSCAN

    Example: Customer segmentation based on buying patterns

    4.2 Association

    • Finds rules or relationships between variables

    • Often used in market basket analysis

    Example:

    Customers who buy bread → also buy butter (with 70% probability)

    Algorithm: Apriori Algorithm

    4.3 Dimensionality Reduction

    • Reduces number of features while keeping important information

    • Helps with visualization and performance

    Common Algorithms:

    • PCA (Principal Component Analysis)

    • t-SNE

    Example: Reduce 50 features of images to 2 dimensions for visualization

Next ❯