Ignoring the tuple: This is usually done when the class label is missing (assuming the mining task
involves classi¯cation or description). This method is not very e®ective unless the tuple contains several
attributes with missing values. It is especially poor when the percentage of missing values per attribute
varies considerably.