A multi-dimensional visualization method combining MDS and SVM

Human can only feel spatial information within three-dimensional space. However, there are more than three attributes in economic statistical data and other data sets generally. When studying the inherent structural characteristics of these data such as clustering and distribution, researchers need to reduce multi-dimensional information to three-dimensional space or less to achieve multi-dimensional visualization. There are many dimension reduction methods, whose results are different from each other because of different mathematics theories and application ranges. In the paper, authors analyze economic statistical data of Sichuan province in 2007 by using Multidimensional Scaling (MDS) which is a nonlinear method and Support Vector Machines (SVM) which is a supervised classification method. The classification result of MDS is consistent with the status of economic development of Sichuan in general, but details of the result cannot be verified itself; the output results of SVM by selecting different kernel functions are very similar to the classification result of MDS, which can validate these results. And considering the advantages and the solid mathematical theory, authors believe that the combination of these two methods is scientific.