FR-Tree: A novel rare association rule for big data problem

摘要

In some situations, finding the rare association rule is of higher importance than the frequent itemset. Unique rules represent rare cases, activities, or events in real-world applications. It is essential to extract exceptional critical activity from vast routine data. This paper proposes a new algorithm called FR-Tree to mine the association rules and produce essential rules. This work aims to demonstrate that this algorithm is suitable for extracting rare association rules with high confidence. The proposed algorithm generates, filters, and classifies the all-important rules, either frequent or rare. The rare rules were produced without needing to set an additional threshold. Therefore, the proposed algorithm has an advantage incomparable with the other rare association rule techniques. The generated rules were tested using well-known datasets, and the performance was compared with the other rare association rule techniques. The results proved that our method outperformed the existing rare association rule techniques.