Clustering and Approximate Identification of Frequent Item Sets

All Papers

We propose an algorithm that computes an approximation of the set of frequent item sets by using the bit sequence representation of the associations between items and transactions. The algorithm is obtained by modifying a hierarchical agglomerative clustering algorithm and takes advantage of the speed that bit operations afford. The algorithm offers a very significant speed advantage over standard implementations of the Apriori technique and, under certain conditions, recovers the preponderant part of the frequent item sets.