One number is rarely enough to judge a model.
Direct answer
Accuracy is the share of all predictions that are correct. Precision is the share of predicted positives that are truly positive. Recall is the share of actual positives that the model found. Choose the metric by cost: use recall when missing a case is costly and precision when false alarms are costly.
Example
A fraud model checks 1,000 transactions. 10 are fraud. It flags 8, of which 6 are really fraud.
- Precision = 6 ÷ 8 = 75 percent
- Recall = 6 ÷ 10 = 60 percent
A model that never flags fraud would have 99 percent accuracy and 0 percent recall.
When to use which
| Situation | Focus |
|---|---|
| Medical screening | recall |
| Spam filter (avoid losing real mail) | precision |
| Balanced need | F1 |
Where Teyro fits
Teyro's Machine Learning for Beginners course turns this into short lessons you can play in a few minutes a day. It has 200 lessons, and you can start free.
The bottom line
Pick the metric that matches the real-world cost of mistakes.
