How to calculate precision and recall?
precision = true positive/predicted condition positive
recall = true positive/condition positive
How to calculate diagnostic odds ratio and F1-score?
What is a confusion matrix?
describes accuracy for a classifier
n x n matrix (true condition vs. predicted condition), where n is the number of classes
reveals specific issues, classifiying between two classes, global accuracy does not does not
How to compute accuracy?
accuracy = true positives + true negatives / total population
Why is observer variability a problem in machine learning?
segmentation and classification based on manual labeling -> subjective
inter- and intra-observer variablity often larger than expected
Zuletzt geändertvor 18 Stunden
