What does the C parameter control in a soft-margin SVM?
PICTURE THIS: DATA SPLIT
Fit on train, tune on val, report on test once.
Simple meaning
C trades off margin width against training mistakes.
WHY — SVM instead of guessing?
Why interviewers care about SVM:
who only read docs from people who shipped.
and tied to AI / ML work.
Name the idea, why it exists, then one short example.
End with when you use it and one common pitfall.
STEPS — What happens step by step?
Before you speak the answer, walk the interviewer through these steps:
- 1C trades off margin
width against training mistakes.
- 2Large C punishes violations
hard and can overfit.
- 3Small C allows a
wider margin and more slack, which often generalizes better on noisy labels.
- 4Give an example
One tiny concrete case you can say aloud.
- 5Common mistake
What juniors usually get wrong.
- 6Close
When you pick this over the alternative.
EXAMPLE — See it in action
Here's a short line you can speak, broken into clear beats:
Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.
Key takeaway
C trades off margin width against training mistakes. Large C punishes violations hard and can overfit.