My Machine learning model have accuracy=80% and precision=85%
what is the reason behind this?
or is it stable or acceptable model?
Hi,
Is this precision is a weighted average precision?
In general, precision is independent of accuracy. The algorithm can be precise but inaccurate, and it also can be accurate but imprecise. This depns on the performance of the algorithm. For more information, have a look at this link:
https://en.wikipedia.org/wiki/Precision_and_recall
Instead of precision, you may need to consider another technique, which is "F-Measure".
HTH.
Dr. Samer Sarsam
Yes this is average precision, but i want to know is the model stable or acceptable?
or why should i use recall?
Precision tells you how accurate you are in predicting positives. With accuracy being low, did you check if recall is acceptable or not. You might have relatively higher false negatives. In general, it is acceptable as long as excess False negatives do not add significant cost.
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning