Suppose the random sample used for each tree in the previous random forest is decreased from 1000 rows to 500 rows. Circle ALL that are true for the ensemble classifier.
a) Variance of predictions are expected to decrease
b) Bias of predictions are expected to decrease
c) Each base classifier is expected to have more out of bag rows of data
d) Generalization error is expected to decrease