noun
Unsupervised learning is a machine learning technique where algorithms identify patterns and relationships in data without predefined labels or human guidance. Unlike supervised learning, which relies on tagged examples, unsupervised learning works with raw, unlabeled data, making it useful for tasks like clustering, anomaly detection, and dimensionality reduction.
It typically uses algorithms like k-means clustering or autoencoders to group similar data points or extract features. This approach is particularly valuable in scenarios where labeling data is expensive or impractical, such as analyzing large datasets in genomics or customer behavior analysis.
For developers, unsupervised learning reduces the need for manual data labeling, cutting costs and time. However, it requires careful preprocessing and validation, as results can be harder to interpret. For example, a retail company might use it to segment customers into groups based on purchase history, improving targeted marketing without needing to label each transaction.
A streaming service uses unsupervised learning to group users into clusters based on watching habits, then recommends shows tailored to each cluster.
In other languages
See also: Supervised learning