Hello, in our project we want to compare the classification accuracy for different subsets of data. Is it necessary to use the same classification model for all the subsets to make a relevant comparison (for example, kNN)? If yes, do the hyperparameters have to be identical for all the subsets (for example, the same number of neighbors)?
Hi, I guess it depends on why you are studying the accuracy on the subsets. What do you want to study? Are the subsets random or split according to some specific criterion?
Comparing accuracy for different data subsets
Hello, in our project we want to compare the classification accuracy for different subsets of data. Is it necessary to use the same classification model for all the subsets to make a relevant comparison (for example, kNN)? If yes, do the hyperparameters have to be identical for all the subsets (for example, the same number of neighbors)?
Thank you!
Hi, I guess it depends on why you are studying the accuracy on the subsets. What do you want to study? Are the subsets random or split according to some specific criterion?
Hello, the different subsets differ in the brain area they were collected from, we would like to see from which area we can make the best prediction.
Add comment