Research on A Clustering Analysis Algorithm Facing the Complex Fundamental Data Prepared

The globalization of the world economy has accelerated the pace of global cooperation, information has become a modern enterprise to adapt to competition to a powerful weapon in the world. Manual labor in the manufacturing system has been greatly reduced due to the wide application of computer technology, automatic control technology, and artificial intelligence technology, modern production environment, the role of the people from directly involved in machining operations gradually evolved into the machinery, equipment, etc. supervision, control, maintenance and so on. The changing nature of work in the human manufacturing system, asking people to pay more attention to the information received, and the decision-making process. In addition, with the acceleration of the process of modern enterprise information, will inevitably lead to the enterprise has a lot of information, and the timely processing of information on the pressing problems of the modern enterprise solved. Management activities of enterprises but also to promote the gradual shift towards data analysis type. How timely processing of these enterprises have a large amount of data to support effective information management activities to get hidden behind the huge amounts of data with knowledge of the decision-making value, which will help policy makers to make more personalized and targeted decision-making on important process in the face of global competition and the process of information become a modern enterprise. To do this, companies must use technology with powerful data analysis and processing capacity to deal with these vast amounts of complex information in a timely manner. The data mining technology is to solve this problem. Cluster analysis has been in the field of data mining occupies an important position, either as a separate tool to find data distribution, and other data mining analysis algorithm can be used as a pre-processing step. k-means algorithm is a classical clustering algorithm based on partition, its advantage is that the algorithm is simple, fast, fast results for large data sets can, but it also has some limitations, such as the initial focal point difficult to choose the optimal k values ??given in advance, cluster validity is difficult to protect other issues. This thesis is focused on the development direction of the characteristics of modern production as well as modern production management mode, facing the complexity of modern enterprise basic data preparation requirements, use of intelligent analysis method to solve the modern enterprise approach when faced with the complexity of the underlying data, methods ---- GWK -means method. Focus on the problem of the initial centers selected dependencies, and the best value of k in the k-means algorithm to select. The proposed feature weighting method to solve the problem of initial centers dependent; establish a new validity function based on the characteristics of the data attributes, to solve the problem of k value selected, and further test the algorithm results validity. Finally, the the GWK-means method is applied to the actual J. In order to solve the existing enterprise prepared to deal with the problem of the complexity of the underlying data supplier evaluation, customer relationship management, provides a practical and efficient scientific method.