Model Training & Adaptation
Support vector machine
A prediction method using selected learning examples to help divide groups or estimate numbers.
Example
A prediction system separates two categories using selected numerical features.
Why people use it
It offers a prediction method built around separating cases with a useful margin.
What you'll hear
“Where should the dividing boundary go?”
What this means for you
Choose features, kernels and settings through appropriate independent checking.
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
- Can it separate groups with a curved boundary?
- Yes. Suitable versions can handle more complicated divisions than a single straight line.
- Does it always supply a trustworthy percentage chance?
- No. A separate probability-estimation step may be needed, and the result still needs checking.
- Can measurement scales affect its decisions?
- Yes. Some measurements may dominate unless the information is prepared appropriately.