What characterizes unsupervised learning?

Unsupervised learning is characterized by the ability to learn from data without the need for explicitly labeled outputs. This means that an algorithm can analyze the structure or distribution of the data, identify patterns, and make sense of the information without prior knowledge of the categories or groups that may exist within that dataset.

In unsupervised learning, techniques such as clustering and dimensionality reduction are frequently employed. For example, clustering algorithms group similar data points together based on their features, while dimensionality reduction helps in identifying the underlying structure of the data by reducing the number of features while preserving essential information.

The other choices do not align with the fundamental nature of unsupervised learning. Training with labeled data is indicative of supervised learning, whereas predictive modeling typically relies on historical data in supervised contexts to make forecasts. Feature extraction may be a part of both supervised and unsupervised learning but doesn't distinctly characterize unsupervised methods by itself.