Machine learning question(Classifier Evaluation)

Question

2 classifiers evaluation~~~~in bagged committee

I had 2 classifiers~~~~classifier burger and classifier fries~~~

both of them has same error bias~~~with committee size 10~~~

classifier burger has error variance 40~~~found generate non-related model under bagging~~~

classifier fries has error variance 30~~~found generate model with pairwise covariance of 0.3 under bagging~~~~

so which one is better? why?

Answer 1

I would prefer the classifier fries, because the error variance is telling that with a smaller volume of data , it's explaining an entire data set due to which the model predictability will is higher. Having higher variance would largely impact the accuracy of the model.