0

In some papers, I find some strange evaluation metrics that did not give a clear explanation, like average class accuracy and average instance accuracy!

  1. What is the difference between the both? I can not find any explanation about them...

  2. how to calculate them in Python3?

Zeaggler
  • 1
  • 1
  • 1
    Can you provide a link to the sources where you found such terms ? – Nicola Mingotti Aug 24 '19 at 03:00
  • ok, thank you very much at first. 1. https://arxiv.org/abs/1604.03265 Fig.7 and 8; 2.https://uwaterloo.ca/mobile-sensing/sites/ca.mobile-sensing/files/uploads/files/2019-wang-cheng-etal-li-neucom.pdf Table 1 and 2. – Zeaggler Aug 25 '19 at 05:40
  • ohhh, i think i got a clue about what you are talking about. `class` is meant in the statistical sense probably, not in the computer science one. see here. https://stackoverflow.com/questions/39770376/scikit-learn-get-accuracy-scores-for-each-class – Nicola Mingotti Aug 25 '19 at 05:45
  • yeah, so... for average class acc., need to calculate the sum of each class acc, and then divided it by class number, am I right?? and what is the meant of instance or sample? – Zeaggler Aug 25 '19 at 05:59
  • Umm, if in your context they use `instance` and `sample` as synonyms I reccomend you stick to `sample`, that is classic statistics terminology. Then, since this seems more a Statistics question than ComputerScience one I suggest you post it in CrossValidated , that's a very specific forum. https://stats.stackexchange.com/ – Nicola Mingotti Aug 25 '19 at 06:08
  • okok, got it, thank you very very much – Zeaggler Aug 25 '19 at 06:14

0 Answers0