Lesson 2 of 6 6 min · 108 words
Accuracy
Simple, and misleading on imbalanced data.
Accuracy is the fraction of correct predictions: . It is intuitive and fine when classes are balanced and all errors cost the same.
On imbalanced data it misleads badly. If 99% of transactions are legitimate, a model that always says 'legitimate' is 99% accurate — and catches zero fraud.
Always compare to a baseline: the majority-class rate, or a simple model. An accuracy of 92% means little if the baseline is 90%.
Formula to remember
Precision
Of predicted positives, how many are right.
Quick check · +5 XP