Browse Teach Curate

Predictions and Ground-Truth Labels in Supervised Learning

The supervised setup in which a model emits a prediction for an input whose correct answer is already known, so that prediction and label can be compared pair by pair. Distinguishes continuous targets (a number to hit) from categorical targets (a class to pick), which is the split every later error metric is defined against.

jeremy Video

Ground Truth Data in Supervised Machine Learning

Ground truth data constitutes the verified ground reference used to supervise machine learning algorithms during training, validation, and testing phases within supervised learning frameworks. This t…