The Data Mining Research For Several Networks

Data mining is the process of selecting, exploring and modeling through large amounts of data, and found the rules and relationship unknown in advance, in the end the owner of the database can get the useful results. The data mining's results have been widely used in finance, logistics, healthcare, retail, manufacturing, engineering and science industry, and bring them a huge profit.In modern society network is everywhere, and there are specific and abstract. In short, it affects people's lives in every area of the society. So the content of this paper is the applications of data mining in a variety of network, and it is included in the following:First of all, we study the easy network formed on the map. We use the GPS data of the private car owners in the database of ChinaGPS, and successfully restore the network formed by parking spots of the private owners. We put forward our own data mining method through the relationship between the network points. We propose the analysis methods which are all based on GPS latitude and longitude, such as the criteria of judging home, determining the points of interest and the index of consumer interest of the cars'owners and so on. We verify the correctness of selecting GPS latitude and longitude as the characteristics through the center clustering by SAS software. I explore the private owners'consume tendency successfully, so it bring a new profit growth point for the company.Secondly, we study the logistics network formed from the tobacco logistics project. According to the customer's actual needs, it formed a complex logistics network of multi-warehouse and multi-delivery points. We analyze the actual situation, and design the mathematical model with complexity of constraints and make up the complex networks by genetic algorithm. At the same time we modify the optimization model real-time based on the customer's feedback, so that the mathematical model is completely in line with customer's need and objective conditions. So we applied the data mining method in the complex and abstract networks successfully.Finally, we have a research on converting the complex topology of the network to time series initially. The visibility graph method is a fast calculation method of changing time series into the complex networks. In this paper we propose the de-visibility graph method using the idea of the visibility graph method, and put the complex network into the two-dimensional time series which is intuitive and easy to understand, so we take the next step in the direction of putting the complex networks into time series.