Artificial and Real dataset with different concept drift type and imbalance ratio

The zip file contains both real and artificial datasets having different concept drift type and imbalance ratio. The real datasets are usually used as a benchmark for concept drift problems and they are PAKDD, Elec2, and KDD99. The artificial datasets are divided into 3 folders, based on the concept drift type: P(y), P(X|y), P(y|X). Each folder contains 16 datasets, 8 generated through the SINE1 generator and 8 with the SEA generator. If their name starts with sine or sea, means that they contain abrupt concept drift at the 50000˚ instance. Otherwise, if their name starts with sineG or seaG, means that they contain gradual concept drift starting at the 45000˚ instance and ending at the 55000˚ instance. Moreover, the numbers that follow state the imbalance ratio.