Metrics in Machine Learning

Updated on: 25 Aug, 2020 || Topic: ML - Metrics in Machine Learning

List of metrics in machine learning according to task

Classification

Segmentation

Regression

Medical

FROC

The free-response ROC (FROC) curve is the most commonly used metric for abnormality detection in medical imaging. Receiver Operating Characteristic (ROC) and Free‐Response Operating Characteristic (FROC) methods are used to assess the accuracy of radiological imaging systems. ROC methods analyze an observer’s confidence that an abnormality is or is not present, whereas FROC methods additionally require the observer to locate abnormalities. Typically, ROC and FROC methods are applied to answer different research questions: sometimes ROC is the most appropriate and sometimes FROC. However, some believe that ROC is usually either equivalent or inferior and FROC is preferred over ROC for all research questions. This is the topic debated in this month’s Point/Counterpoint.

References

FROC


Author: Jitender Singh Virk
virksaab.github.io