Beginner · Data

Sampling (data)

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

TL;DR. Selecting a subset of records from a larger dataset for training, evaluation, or analysis.

Technical Definition

Selecting a subset of records from a larger dataset for training, evaluation, or analysis.

How it works

Random sampling preserves overall distribution; stratified sampling preserves per-class proportions; importance sampling weights examples by how informative they are. Wrong sampling introduces bias before a single model is trained.

Related Concepts

• Cross-Validation — A technique that evaluates model performance by training and testing on different subsets of the data in rotation.
• Dataset — An organized collection of examples used to train, validate, or test a model.
• Data Imbalance — When some classes or groups are vastly more frequent in the dataset than others.