Technology On Fast Visualization Of Large Scatter Plot Matrix

In recent decades, data mining appears and gets more and more attentions and becomes a new field of computer science. People get useful information from large scale data by using data mining technology. For making the process and results of knowledge discovering and the data mining easier to show, people use and develop data visualization technology. Scatter plot matrix data visualization technology shows the relationship between data by a graphical method and at the same time it can reduce the dimensions of data. So scatter plot matrix becomes one of the most useful data visualization methods. Many data mining software have scatter plot matrix visualization function, but as the scale of data becomes large, the result of data visualization is not clear, lack of interactivity and has low speed, which are the problems this paper will solve.This paper focuses to some problems of data visualization in the date mining systems, such as the speed of data visualization becomes very low along with the scale of data becomes large; the result of data visualization is not clear when the data size becomes large; the data visualization function in present software does not support the interaction between user and system. In this paper I will propose a data structure N-Box Tree to achieve fast data visualization and interaction between user and system.This paper introduces the idea, construction, operations of N-Box Tree, and applies the N-Box Tree data structure to data visualization technology. N-Box Tree data structure can improve the speed of data visualization and interactive operations between user and system. At same time this paper analyzes the performance of N-box Tree by the experimentation and proves that the speed of data visualization can be improved by using N-Box tree data structure.At the same time this paper introduces two methods to improve the performance of data visualization. They are sample technology and cache technology, and get good result by the experimentation.At last, I develop a scatter plot visualization system based on AIBox system. The system includes four modules: data inputting module, data preprocessing module, data sampling module, data visualization module. The implement of this visualization system applies an auxiliary module for data mining system.