Research On A Clustering Analysis Algorithm Facing The Complex Fundamental Data Prepared

The world economy globalization accelerates the pace of global cooperation, and information becomes the powerful weapons for modern enterprise meeting competition. In modern production environment, due to computer technology, automatic control technology and the wide application of artificial intelligence technology in manufacturing system, the physical labor has declined, and the person's role is gradually going from directly involving in processing operation to supervision, control and maintenance of machines and equipments. The change of human nature's work in manufacturing system is asking people to pay more attention to the information receiver, processing and the decision-making process.In addition, the speeding up of modern enterprise informationization advancement inevitably leads the enterprise to make large information, and to process information of the modern enterprise in time becomes the urgent problems urgently to be solved. At the same time it prompts the model enterprise management activities gradually tending to data analysis. Therefore, how to put these large numbers of data information to support timely treatment of enterprise management activities so as to get an effective information in these hidden behind massive data with decision value knowledge and favor the decision makers to make more individuation and targeted decision-making, becomes the important process modern enterprises facing the global competition and informationization. And enterprises have to use technical data processing which has strong ability of timely processing of these mass complex information. And then data mining technology is solved this problem for enterprises.Clustering analysis has been occupying an important position in the area of data mining, it can find data distribution as a separate tool, also can serve as a preprocessing step for other data mining algorithm. K-means algorithm is a kind of classic clustering algorithm based on division, and its advantage is simple, fast, and it can obtain the result for large data sets quickly. It also has certain limitation, such as initial center difficult to choose the optimal, a given k value, clustering validity problem difficult to ensure, etc.According to the characteristics of modern manufacturing and the development direction of modern enterprise for modern production management mode, this thesis mainly sets to solving complex based data processing problem by using intelligent analysis method for modern enterprise, which is named GWK-means method. It discusses emphatically initial center choice and selection for optimum k value problem of K - means algorithm, which puts forward to solving the clustering result dependence on initial center problem using weighted feature and sets up a new validity function based on data attributes to solve k value selection problem. And then this essay tests the effectiveness of this method.Finally, GWK-means method is applied to solve the practice existing problem. And it is provided as a practical and efficient scientific method for complex based data prepared including supplier evaluation and customer relationship management for J Company.