Unsupervised Learning

A type of machine learning where an algorithm learns to identify patterns in data without being given the correct answer. Unsupervised learning is most commonly done after supervised learning so that the model is trained already to understand the patterns and can build on this existing knowledge.

How It Works

1. Input raw data in to a model pre-trained through supervised training
2. Find similarities and patterns
   • Uses mathematics like distance measures, probability and distributions, i.e. how similar or different each data point is compared to the others
3. Detect anomalies.
4. Group and simplify the data
5. Output advanced patterns

Key Techniques

• Clustering - grouping similar data points together (e.g. normal vs suspicious login attempts)
• Dimensionality Reduction - Simplifying large datasets by reducing the number of variables while retaining a working pattern