Application Of Principal Component Analysis And Clustering In Science And Technology Data Analysis

After the reform and opening up, the development of scientific and technical gradually increased. Aspects of the gradual enhancement of science and technology system reform continued to deepen, and the national innovation system was constantly moving forward steadily, which followed the law of economic development, and technology of their own development law. The current pattern of technology development met the needs of the market and market demand-oriented is slowly developing.Configuration of scientific and technological resources is being further optimized process, and the configuration of market resources is under the guiding role of the market rules of spontaneous complete. Many national resource projects are achieved through competition, and the market is the result of competition survival of the fittest,then it is the market itself.By studying the China Statistical Yearbook of scientific and technical data related, we obtain statistical data from country’s 31 administrative units, which includes the number of patent applications, the newly developed several projects,product or process innovation activities of the number of enterprises, surveying benchmark results, and other 51 Statistics variable.In this paper, we make data analysis for the statistical data 31 administrative units, and it mainly to complete two things.(1)Do principal component analysis on sub-regional scientific and technical data.After analyzing "China Statistical Yearbook" equatorial area of science and technology-related statistics, we obtain 51 properties of the variable from 31 administrative units, but it has higher dimension, so we make principal component analysis for the data.Principal component analysis is based on multiple attributes into a small number of process variables of several main components. In the process of the principal component analysis, it achieves two objectives. The first aim is to reduce the dimensionality of a simplified data structure; and the second aim is that a data matrix consisting of a few main components can also reflect most of the data. Principal component analysis can be said to be a process variable properties concentrated.After principal component analysis, the 51 attributes variable dimensionality reduces, with eight main components to represent. By calculating the main component score, we get 31 administrative units of science and technology related to the total score of 31 administrative units. We rank the score and analyze the different administrative units of the technical differences science.(2) Do clustering analysis for scientific and technical data by region.Cluster analysis is an important means of knowledge. It is the use of "like attracts like" thinking to classify data. The similar objects are in the same class, and the characteristics of different objects are in different classes.The results of principal component statistics 31 administrative units are subjected to K-means clustering analysis, and we get three clustering results, in turn defined as "the science and technology developed areas", "the general area of science and technology" and "underdeveloped areas of science and technology ", and we make the relevant reference data analysis for China’s economy.