Abstract: In batch learning, it is commonly assumed that samples are independent and identically distributed (i.i.d.). However, this assumption does not hold in dynamic environments where data streams ...