Design And Implementation Of Enterprise Competitive Intelligence System

With the rapid development of modern computer and network technology, Internet is becoming an alternative to traditional access to information and an efficient way to quickly access to competitive intelligence information. However,how to discover valuable intelligence information and how to corporate decision-making to provide timely, accurate and effective business intelligence are importance for enterprise development. Among them, the enterprise competitive intelligence system is an effective way to obtain competitive intelligence. How to build efficient enterprise competitive intelligence system to meet the needs of enterprise is a meaningful subject.This paper focused on Enterprise Competitive Intelligence System(CIS) based on Big Data. First, the research state of competitive intelligence system was reviewed; then, economic and technical feasibility was discussed, and some construction strategies of CIS based big data were analyzed. Lastly, a prototype of CIS was developed to test and verify the model and tactics put forward in this paper.This research includes three parts:(1)The theories of Competitive Intelligence and Competitive Intelligence Systems was Elaborated. then, the concepts and characteristics of big data were introduced. Two main techniques, Hadoop Distributed File System and NoSQL database were emphatically analyzed; the Challenges and issues faced to CIS were summarized.(2)The models of CIS based on Bid data was proposed. the feasibility of its implementation and system functions was discussed.Three subsystems, intelligence gathering subsystem, intelligence analysis subsystem and intelligence services subsystem, were Emphatically analyzed.(3)A prototype system of enterprise competitive intelligence was developed. the paper focused on method of information collector. Algorithm TF-IDF in Hadoop environment was tested, which experimental results show that under the conditions of large-scale data, Hadoop implementation based on TF-IDF is calculated to be significantly due to the traditional approach.