|
Data mining through large amounts of data selection, exploration and modeling process, discovered a previously unknown rules and contact database owners, clear and useful results. The research results have been widely used in the industries of financial services, logistics, healthcare, retail, manufacturing, engineering and science, they bring a huge amount of profits and considerable development. Networks ubiquitous in modern society, a specific, abstract and affect people's lives, in short, in all areas of society. Therefore, this project will be data mining technology in a variety of network applications, this study mainly include the following aspects: First, we study the network formed on the map. With the private owners GPS Navigation Technology Co., Ltd. of Shenzhen SEG database data, we successfully restore the network formed by the private owners of parking spots. Then we propose the method of data mining, the analysis of the relationship between the network points decision shelter guidelines to determine the points of interest guidelines and owners of consumer interest in index-based method of analysis of the characteristic quantities of the GPS latitude and longitude. And SAS software running center clustering algorithm to verify the correctness of the selected GPS latitude and longitude as characteristic quantities. By analyzing the successful exploration of the propensity to consume of private car owners to bring new profit growth point for the company. Second, our the tobacco logistics project logistics network. According to the actual needs of the customer, the formation of a multi-warehouse, multi-delivery point, complex logistics networks for a variety of products. By analyzing the actual situation, I design a mathematical model with complex constraints and the calculation of complex network by Genetic Algorithms. Based on customer feedback, real-time modify the optimization mathematical model, the mathematical model to fully meet customer needs and objective conditions. Thus, the success of the data mining methods applied to a relatively complex and abstract network. Finally, we converted to a time series of complex topology network problem has conducted a preliminary study. Visibility graph is a time series into a complex network fast calculation method. Paper, I take advantage of the can view law thought, put forward the inverse can view law, complex network topology conversion time series inverse process, made the preliminary research results in the corresponding relations between the two.
|