This is a synthetic data set of 40 features, representing objects from two classes (labeled as 0 or 1). The training set has 1000 samples and the testing set has 9000.