Unsupervised Learning

Types of ML Algorithm

This type of learning algorithm learns to identify patterns or structure in a set of features (X) without any corresponding output labels (Y).

The goal of unsupervised learning is to discover hidden relationships or insights in the data, and it can be useful for tasks like clustering, dimensionality reduction, and anomaly detection.

Unsupervised Learning is done on unlabeled data.

Unsupervised learning algorithms can be divided into several major types:

  1. Clustering

  2. Dimensionality Reduction

  3. Anomaly Detection

  4. Association Rule Mining


These algorithms group similar examples in the input data based on patterns they find.

Example: Identifying different groups of customers based on their purchasing behavior, such as grouping customers who tend to buy electronics, customers who tend to buy clothes, and customers who tend to buy food.


  1. K-means Clustering

  2. Hierarchical Clustering


Dimensionality Reduction

These algorithms try to find a lower-dimensional representation of the data that still preserves as much information as possible.

Example: Visualizing high-dimensional data such as images or audio data in a lower-dimensional space that humans can understand.


  1. Principal Component Analysis (PCA)

  2. t-SNE

  3. Autoencoders.

Anomaly Detection

These algorithms identify unusual or suspicious data points that don't fit with the rest of the data.

Example: Identifying fraudulent credit card transactions based on patterns in the data.


  1. K-means Clustering

  2. Gaussian Mixture Models

  3. Isolation Forest

Association Rule Mining

These algorithms find frequent co-occurrences of items in transactional data.

Example: Finding patterns in online browsing behavior to make product recommendations to customers.


  1. Apriori

  2. FP-growth

Overall, unsupervised learning algorithms can be used to discover patterns or structure in data without any explicit guidance or labels.

Last updated