An Approach for Mining Frequent Sequences in Distributed Environment

摘要

Discovering meaningful information from data is what is known as data mining. By meaningful information we mean many mining problems such as association rules, frequent sequences, or clusters. So far, attention in a data mining process has always focused on extracting information from centralized data or data physically located at one single site. However, there are some applications in which data distribution in different sites is a need or mandatory. In this case, the process embraces additional complexity. One aspect of complexity concerns with how to adapt or innovate mining algorithms to extract information from distributed data rather than central one. Another problem, concerns with setting a mechanism for communication involved in the distributed environment without incurring significant overhead. This paper is a preliminary study for mining distributed sequences in light of current mining algorithms. Based on detailed comparisons of algorithms in sequences mining area, we propose a distributed approach to follow for implementation of sequences discovery.