Comparison with SLIQ

• Difference in Attribute list structure: In SLIQ, class label is stored outside the attribute list structure
• Another list called Class List is used that is indexed by rowid and it's record entry contains class label and a pointer to a node of decision tree that indicates the node to which current data belongs
• The class list has to stay in memory because this list is randomly accessed and updated. It limits the size of training data that SLIQ can handle.
• SPRINT can handle very large data easily and classifies them efficiently