Machine learning is a broad branch of data science. Also referred to as statistical learning.
Includes [[supervised learning]], [[unsupervised learning]], and [[reinforcement learning]].
Machine learning tasks include model evaluation and model selection.
Read the latest in [Papers With Code](https://paperswithcode.com/)
## model evaluation
- holdout
- [[k-fold cross-validation]]
- [[bootstrap]]
confusion matrix
sensitivity, specificity, precision, accuracy, F1
ROC, AUC-ROC
## model selection
(mean/relative) absolute error
(mean/relative) square error
t-test for model selection
$t = \frac{\bar {err}(M_1) - \bar {err} (M_2)}{\sqrt{var(M_1 - M_2)/k}}$
degrees of freedom is $k-1$ in [[k-fold cross-validation]].