Dynamic weighted sequential pattern mining for USN system
Pilsun Choi, Bu-Hyun Hwang · 2017
Sequential pattern mining is a technique which finds out frequent patterns from the data set with a time attribute. Dynamic weighted sequential pattern mining can be applied to a computing environment that is constantly changed according to time period and it can be applied to a variety of environments with changes of weight value. In this paper, we propose a new sequential pattern mining method to discover frequent sequential patterns by applying the dynamic weight. This method reduces the number of candidate patterns by dynamic weight according to the relative time sequence. This also decreases the memory usage and processing time compared with existing methods. And we show the importance of processing USN (Ubiquitous Sensor Network) stream data with hash map structure and the method of renewing current frequent pattern list.