Mining Approximate Time-Interval Sequential Pattern in Data Stream

Provided by: AICIT
Topic: Big Data
Format: PDF
Most existing sequential pattern mining algorithms are hard to discover long meaningful time-interval sequential patterns in data stream. In this paper, the authors present a new bitmap-based algorithm of mining approximate time-interval sequential pattern in data stream called BMATS, which is based on binary bit counting and multiple time-interval sequence alignment. They translate the whole sequence database into bitmap and all operations into binary bit counting to improve the efficiency of their algorithm. Moreover, instead of exact matching, they use multiple sequence alignment to discover long meaningful time-interval sequential patterns.

Find By Topic