KDD98

Dataset KDD98 challenge: https://kdd.ics.uci.edu/databases/kddcup98/kddcup98.html The goal is to estimate the return from a direct mailing in order to maximize donation profits. This dataset represents problem of binary classification - whether there was a response to mailing. For this version, the target was correctly encoded as a binary factor. The features 'HPHONE_D', 'MHUC2', 'INCOME', 'WEALTH1', 'WEALTH2' were recoded as nominal factor variables and the constant feature 'RFA_2R' was removed from the dataset. For this version, the majority class was downsampled to 40% of the original size. Unused factor levels were dropped.