Skip to main content

Unsupervised Learning

Updated May 04, 2023 ·

Overview​

Unsupervised learning is similar to supervised learning but without a target column. It learns from the dataset to find patterns, providing insights without needing to know much about the data beforehand.

  • Learns from the datasets, trying to detect patterns.
  • Clustering, anomaly detection, and association.

Clustering​

Clustering identifies groups in your dataset where observations in the same group share more similarities with each other than with those in other groups.

  • Clustering of animals by breed
  • Clustering of animals by color

Below are some examples of clustering models:

  • K-Means

    • Requires specifying the number of clusters.

  • DBSCAN

    • Density-based spatial clustering of applications with noise.
    • Defines clusters based on density
    • Doesn't need the number of clusters defined in advance.
    • Instead, it requires you to define what constitutes a cluster.

Anomaly Detection​

Anomaly detection focuses on identifying outliers, which are observations that significantly differ from others.

  • Detecting Outliers

    • Example: A point that stands out from a cluster of points, such as a sum total that wasn't removed.

  • Removing Outliers

    • Easier to identify with fewer dimensions, but challenging with many dimensions, requiring algorithms.

  • Anomaly Detection Use Cases

    • Finding errors, identifying faulty devices, detecting fraud, or recognizing patients with unexpected disease resistance.

Association​

Association involves finding relationships between observations, often used in market basket analysis to determine which items are bought together.

  • Finding events that happen together.
  • Often used in market basket analysis - "which object are bought together?"
  • People who buy jam often buy bread, beer with peanuts, and wine with cheese.