Write functions to obtain performance metrics

cytodata / single-cell-classifier

Software to classify biological phenotypes from single-cell image-based morphology profiles

Creative Commons Attribution 4.0 International

7 stars 7 forks source link

Write functions to obtain performance metrics #26

Closed gwaybio closed 4 years ago

gwaybio commented 4 years ago

One output for all models (using both EBImage and raw image features) will be predicted single cell classes. We should write a series of functions that each script can import that take in predictions and true class labels and output a series of performance results (including performance metrics, confusion matrices, etc.)

wagenrace commented 4 years ago

Can this be considered done?

gwaybio commented 4 years ago

lets keep the issue open - there are many more useful performance metrics and functions that we should consider implementing

Examples:

Code to calculate sensitivity, specificity, precision, recall
Code to plot ROC and precision recall curves

wagenrace commented 4 years ago

Would it be better to make those into separate clearly defined issues?

gwaybio commented 4 years ago

if it's helpful, go for it! As a maintainer, you are welcome to design this repository as you see fit