Bagging is useful to reduce the bias if algorithms are high bias

In Bagging, models are trained in parallel 

In Bagging, bootstrap is used to sample from features

Bagging is useful to reduce the variance with big data