Wrapper Method

Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary

A feature selection approach that scores candidate feature subsets by training the model on each one and keeping the subset that performs best.

What is a Wrapper Method?

A wrapper method is a feature selection technique that evaluates feature subsets by training the actual model on each candidate and measuring its performance. The model “wraps around” the search: whichever subset scores best on held-out data wins.

It is one of the three standard approaches to feature selection, alongside filter methods (which score features statistically, without training a model) and embedded methods (which select features during training itself).

How Wrapper Methods Work

A wrapper method needs three parts: a search strategy that proposes subsets, a model to train on each proposal, and a performance metric to compare them. The search continues until adding or removing features stops improving the score.

The common search strategies:

  1. Forward selection starts with no features and adds the one that improves performance most at each step.
  2. Backward elimination starts with all features and removes the least useful one at each step.
  3. Genetic algorithms evolve a population of feature subsets, combining and mutating the strongest candidates.

The trade-off against filter methods is cost versus fidelity. A wrapper trains the model dozens or hundreds of times, which is expensive – but the score reflects how the features perform in that exact model, interactions included, rather than in isolation.

Example of a Wrapper Method

A telecom team predicts customer churn from 40 available features. Forward selection starts with an empty set and tries each feature alone; “months since last contract renewal” gives the best single-feature accuracy, so it stays.

The search repeats. With renewal fixed, “customer service calls” adds the most accuracy; then “monthly charges.” By the fourth round, no remaining feature improves the validation score, and the search stops.

The final model uses 3 features instead of 40. It trains faster, and cross-validation shows less overfitting than the full-featured version.

Related AI terms: Feature Selection · Feature Engineering · Information Gain · Cross-Validation · Overfitting

Did you like the Wrapper Method gist?

Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.

Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About

Read the Governor's Letter

Stay ahead with Governor's Letter, the newsletter delivering expert insights, AI updates, and curated knowledge directly to your inbox.

By subscribing to the Governor's Letter, you consent to receive emails from AI Guv.
We respect your privacy - read our Privacy Policy to learn how we protect your information.

Browse All AI Terms A–Z

Every term in the dictionary, in alphabetical order. Jump to a letter or scroll the full list.

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z