Optimized Distributed Data Mining using Multi Agent System

Abstract

Data Mining is an emerging field which has a great emphasis on the concept of finding patterns and trends from large
quantities of information. Data Mining is performed on Data Warehouse which is simply large amount of stored temporal
information which is stored at a central site. Then it is accessed and algorithms are run on them to extract useful modules
of information from that data. However there are many complex systems available in which single data mining is not
feasible, for that Distributed Data Mining is necessary. Distributed Data Mining (DDM) is widely used in commercial
world to get effective results faster than simple or single data mining (SDM) technique [1]. DDM simply uses divide and
conquer technique in wider terms in which large amount of data are divided into small chunks and passed onto different
systems where those individual systems run algorithm and generate their own results. After that all the results are
combined together centrally. This solves the problem of Resource Constraint and allows better resource utilization with
less expense. DDM Architecture with Multi Agent is a system having multiple agents capable of reaching goals which
are very difficult for a single system to achieve.
This paper is the integration of two technologies namely multi-agent system and distributed data mining, also known as
multi agent based distributed data mining, in terms of significance, system overview, existing systems, and research
trends.