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 effective unless the tuple contains several
attributes with missing values. It is especially poor when the percentage of missing values per attribute
varies considerably.