Research Article
Discovering Significant Sequential Patterns in Data Stream by an Efficient Two-Phase Procedure
| Input: T: a tree, H: a header table, and base-item | | : support threshold | | k: user-specified length | | xs: testable support | | Output: candidates | (1) | Candidates = () | (2) | For each item Q in H (with a bottom-up sequence) do | (3) | If H.Q.support > xs then | (4) | base-item = Q base-item | (5) | If |base-item| ≤ k and base-item.support ≥ MinS then | (6) | Copy the base-item to Candidates | (7) | Create a subtree subT and a subheader table subH | (8) | SSPs_Candidates (subT, subH, base-item, k, xs, and ) | (9) | End If | (10) | End If | (11) | Remove Q from H | (12) | End For | (13) | Return Candidates |
|