Model Training & Adaptation
Linear regression
A method that predicts a number by adding together clues, giving each a particular amount of influence.
Example
A simple AI system estimates demand from price and advertising spend.
Why people use it
It offers a relatively simple way to estimate a number from other measurements.
What you'll hear
“How does demand tend to change with price?”
What this means for you
Check whether a straight-line relationship is reasonable for the intended use.
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 the method make predictions outside the examples it learned from?
- Yes, but those predictions can be unreliable if the earlier pattern stops holding beyond that range.
- Can it make sensible guesses far beyond past examples?
- Not reliably. Extending a fitted line too far can lead to unrealistic estimates.
- Can it use several pieces of information?
- Yes. The estimate can combine several measurements, not just a single factor.