Thesiss

In this dissertation we propose solutions to four major problems encountered by data stream classification, namely, infinite length, concept-drift, concept-evolution and limited labeled data . Traditional data stream classification techniques address only the infinite length and…