The Cooperation Mechanism of Multi-agent Systems with Respect to Big Data from Customer Relationship Management Aspect

摘要

Unarguably, with the unparalleled emergence of metamorphic utilization of mobile computing gadgets combining with the social networks. Hefty and massive amount of data are unprecedentedly generated within a second. Search engines host diversified streams of information have created unprecedented scattered data. Hence, effective management and the capability to process large-scale data pose an interesting but critical challenge for contemporary business organizations. Substantively, customers are expanding their online footprints extensively, which makes it hard to extract data value through data collection and data mining. Due to the distributed databases embedded based on heterogeneous platforms, business organizations are facing problematic challenges. It becomes urgent research issues to efficiently and effectively conducting data mining mechanisms with respect to massive amount of data to meet the organizational strategic objectives. Evidently, Big Data era has witnessed the rigorous challenges concerning data transferring, integration, and data-processing technologies. Proverbially, the commonly known Intelligent Agents (lAs), as the autonomous entities to direct its actions towards diverse goals in order to satisfy the implicit requirements for high-speed data integration as well as cooperation mechanisms among different heterogeneous databases. Literally, a Multi-Agent System (MAS) can deal with the flexible communication and cooperation among distributed intelligent agents as an information processor. This paper will introduce multi-agent systems and their applications from data mining aspect, followed by the value of data mining from Customer Relationship Management (CRM) aspect. At last, we propose a three-step data-mining model, which can help business organizations to dig out potential value to manage CRM optimally including using K-means to cluster massive data. In addition, we generalize data to focus on relevant attributes via using information gained and information entropy calculation method to make decision trees for extracting potential valuable knowledge purpose.