1

I calculated the accuracy for a model with listwise deletion (0.75) and for a multiple imputation model (0.67).

It seems like the standard model performs better. However, the naive baseline of the standard model is 70% (the majority class is 70% of the observations) while the naive baseline of the imputed dataset is 60%.

Should I compare 0.677 (accuracy of the multiply imputed model) with 0.75 and conclude that it performs worse, or should I conclude that 0.677 compared to 0.6 (naive baseline imputed dataset) is better than 0.749 compared to 0.7 (naive baseline complete cases dataset)?

I hope you can help me!

tessa
  • 11
  • 1
  • The answer would depend on the details of your dataset and the end purpose. But before any of that, do I understand correctly that this is a classification task with a naive bayes? What are the features that you're using and how are you imputing them? – user4718221 Jan 15 '22 at 04:47
  • Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. – Community Jan 25 '22 at 10:45

0 Answers0