13 Jul 2006
Mining time-delayed association from event datasets
Speaker: LOO King Kong
Abstract
We study the problem of finding time-delayed associations among types of events from an event dataset. Based on the properties of time-delayed associations, we present an algorithm for the problem. Starting from "short" associations, the algorithm repeatedly form and evaluate "longer" associations from those found to be frequent in due course.
Read the Presentation
Slides...
|