Beginner · Data

Training Data

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. The portion of a dataset used to fit a model's parameters.

Technical Definition

The portion of a dataset used to fit a model's parameters.

How it works

Training data is what the model literally sees and learns from. It typically makes up 70–90% of the full dataset. Biases or errors in training data propagate to the model — 'garbage in, garbage out' is the iron law of ML.

Related Concepts

• Data Augmentation — Techniques that artificially expand training datasets by applying transformations to existing samples.
• Dataset — An organized collection of examples used to train, validate, or test a model.
• Validation Data — A held-out subset used to tune hyperparameters and detect overfitting during training.