Modeling support changes in streaming item sets

Abstract

Monitoring of changes in trends discovered from data streams has been recognized as an important problem in view of the growing number of applications such as network flow analysis, e-business, stock market analysis etc. Monitoring of changes in trends necessitates periodic mining of the stream, which is a challenging task because of the high-speed, high-volume, only-one-look characteristics of the data streams. User subjectivity in monitoring of the changes adds to the complexity of the problem. This study addresses the problem of modeling changes in applications that monitor frequency behavior of item sets and proposes an approach to integrate user subjectivity with the modeling task. In our approach, user subjectivity is captured through a three-stage strategy for focusing on item sets, which are of current interest to the user. Modeling of changes in their frequency (support) behavior is proposed, taking into account the user's notion of change. Metrics to quantify the changes detected in the support behavior of the interesting item sets in the data stream are introduced. This study also proposes an architecture of a user centric SI-MON monitoring system to carry out the desiderata. This system continually mines the support of user-specified “interesting” item sets, and discovers and models changes in their frequency behavior.