3. Mining weighted erasable patterns over sliding window-based data streams
In this section, we propose a weighted erasable pattern mining algorithm for processing sliding window-based dynamic data streams. At first, we introduce several important definitions of traditional erasable pattern mining algorithms [36], [29] and [32] and provide the corresponding empirical examples in the Preliminaries section. Thereafter, with the overall architecture of the proposed algorithm, we describe our own data structures proposed in this paper and various mining techniques including pattern extracting and pruning methods suitable for the weighted erasable pattern mining over sliding window-based data streams.