Frequent Itemsets Mining Algorithm for Uncertain Data Streams Based on Triangular Matrix

Abstract

Aiming at the problem of frequent itemsets mining in uncertain data flows, this paper proposes a botm-mine for frequent itemsets mining in uncertain data flows. In this algorithm, trigonometric matrix, queue and frequent item set tree are used to construct the profile structure to store the relevant data flow information of transactions. The support degree of items $1_{-}$ and $2_{-}$ itemsets is efficiently stored in the matrix through matrix. Compared with the transaction matrix, it not only saves space, but also reduces the complexity of computing each support degree, and at the same time has better space-time efficiency.