A hybrid approach to mining frequent sequential patterns

The mining of frequent sequential patterns has been a hot and well studied area—under the broad umbrella of research known as KDD (Knowledge Discovery and Data Mining)— for well over a decade. Yet researchers are still uncovering interesting problems, new algorithms, and ways to improve upon existing methods. In this paper, we marry state-ofthe-art frequent sequential pattern mining algorithms (e.g., SPAM, FOF, PrefixSpan), data structures (e.g., aggregate tree, bitmap), and other tried-and-true methods for candidate generation (e.g., apriori), in an attempt to derive a new algorithm with the best qualities of the aforementioned algorithms. In this paper, we disseminate the new algorithm created, lessons learned, and future work to be done. Categories and Subject Descriptors H.2.8 [Database Management]: Database Applications— Data mining General Terms Algorithms, Design Keywords Projection Database, Frequent Patterns, Apriori, Pattern Growth

• Type: Conference
• Year: 2009
• Where: ACMSE
• Authors: Erich Allen Peterson, Peiyi Tang
• Keywords: ACMSE 2009, Frequent Sequential Pattern, General Terms Algorithms, Pattern Mining Algorithms, Theoretical Computer Science