nlp-qual-space / TODO.md
maxspad's picture
updated TODO
3f11a86
|
raw
history blame
851 Bytes

βœ…- Level of detail = description of context and performance βœ…- Suggestion for "improvement" given βœ…- "written" feedback

  • Link βœ… - How specific was the evaluator in describing the behavior? => exact language of Q1 βœ… get rid of overall model just use component model
  • get rid of true comments

βœ… Medical education requires high-quality written feedback, but evaluating these supervisor narrative comments is time-consuming. The QuAL score has validity evidence for measuring the quality of short comments in this context. We developed a NLP/ML-powered tool to assess written comment quality via the QuAL score with high accuracy. Try it for yourself!

  • include confidence bar

βœ… change out to formal descriptions

Paper

  • zotero refs
  • methods comments
  • flowchart figure
  • final word clouds
  • add within-1 accuracy numbers