Efficient Mining of Generalized Negative Association Rules

Most association rule mining research focuses on finding positive relationships between items. However, many studies in intelligent data analysis indicate that negative association rules are as important as positive ones. Therefore, we propose a method improved upon the traditional negative association rule mining. Our method mainly decreases the huge computing cost of mining negative association rules and reduces most non-interesting negative rules. By using a taxonomy tree that was obtained previously, we can diminish computing costs, through negative interestingness measures, we can quickly extract negative association data from the database.