Model Training & Adaptation
One-hot encoding
Representing categories with separate yes-or-no markers, usually written as ones and zeros.
Example
A color field becomes separate indicators for red, blue and green.
Why people use it
It lets software represent categories without pretending that one is numerically larger than another.
What you'll hear
“Give each color its own yes-or-no marker.”
What this means for you
Plan how unseen categories and large category lists will be handled.
Can you control it?
Developer-only
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Does one-hot encoding imply an order between categories?
- No. It is designed to avoid an unintended numerical ordering.
- Can it create a very wide table?
- Yes. Each possible category needs its own place, so many categories can require substantial space.
- What happens with a previously unseen category?
- The system needs an agreed handling rule because the original set has no marker for it.