I'm running a classifier (logistic regression). The information on my dataset are the following:
dataset size= 279 observations
train size= 233 test size = 56 # of events in train = 31 # of events in test = 8
I think my classifier and results may be affected due to this not equal proportion. Is there any way to avoid bias issues and improve accuracy? What do you personally think of such data?