Visualise Web Usage Mining: Spanning Sequences' Impact on Periodicity Discovery

Abstract

In this paper we present a more effective method to discover the periodicity in web log sequence data which handle missing sequences which may occur during the aggregation process, such as sequences that swing between two periods. On other hands, a sequence may start near the end time of a period where the rest of those sequences appear in next period however, these kinds of issues certainly it will leave its effect of periodicity discovery. Moreover, we incorporated OLAP data cube techniques in the aggregation process in order to handle large generated sequences and visualised the discovered periodic patterns, in order to study its impact on periodicity discovery.