I would like to write my master thesis on how well machine learning evaluation metrics align with statistical estimation properties.
For example, is it possible to use the F1 score to draw conclusions about the unbiasedness (and consistency) of the estimator for the population parameter?
I feel like there is no literature on this topic yet. Has anyone already heard of this issue?
I would really appreciate any response!
Machine learning (AI) uses algorithms, wheres statistics utilises data analytics. There is a lot of information online if you look
Machine-learning can be aligned with any kind of calculation or prediction. You can also use evolutionary algorithms for your purpose. The difference between statistical estimation and ML is that statistical estimation usually uses linear techniques but Ml is more accurate.
Good Luck
Hello Sina, I haven't seen any studies that use the F1 score, but we have just published a few studies on this topic, with focus on cyber risk analytics. Check them out, its published in open access: Article Artificial intelligence and machine learning in dynamic cybe...
Article Dynamic real-time risk analytics of uncontrollable states in...
we have a few more published, but this should be a good (and easy) start.- Badges
- Science topic
- Similar topics
- Psychology
- Emotion