Precision
In plain terms
Of predicted positives, the fraction that are actually correct.
Going deeper
Precision measures how many of the items a model flagged as positive truly are, important when false positives are costly. It answers 'when it says yes, how often is it right?'
Under the hood
Trades off against recall; high precision means few false alarms.
Examples
Spam filtering, search relevance, fraud alerts, evaluation metrics.
Think of it like...
Like a spam filter judged on how rarely it wrongly flags good emails.