A General Effective Framework for Monotony and Tough Constraint Based Sequential Pattern Mining

Fast discovery of sequential patterns in large databases using effective time-indexing

“…When processing maxgap, it only projects the extendable items with respect to the prefix. The experiments in [6] show that CBPSAlgm works more effectively than CCSM, especially for low minimum supports.…”

“…Although DELISP handles the time constraints effectively, the size of the projected databases might accumulate to several times the size of the original database. CBPSAlgm [6] is a projection-based pattern growth method for constraint based sequential pattern mining. It efficiently solves monotony and tough constraints like maximum-gap.…”

“…GSP [25] and DELISP [17] accept most of the constraints (including mingap, maxgap, and sliding time-window). Related algorithms such as prefix-growth [23], cSPADE [29], CCSM [21], and CBPSAlgm [6] allow only one or two constraints; they were not used in the comparisons described below.…”

Efficient strategies for tough aggregate constraint-based sequential pattern mining

“…For example, for MaxGap constraint, the available work includes cSPADE [18], CBPSAlgm [2] and CCSM [10]. However, compared with MaxGap constraint, much less work on the other important classes of tough constraints, especially the tough aggregate constraint, has been done.…”

“…Then a projected database is constructed for every new prefix and we recursively invoke PTAC. 2 Fig. 2 presents the pseudo code of PTAC.…”