Carrying on the Pearson related analysis, the canonical analysis, thesimple regression and multiply regression in SPSS for WINDOWS 12.0 statistics software, we find :1, four regional factors affect the space agglomeration of IT industry markedly, particularly for capital resources and infrastructures.

经使用SPSS for WINDOWS 12.0统计软件进行Pearson系数相关、典型相关和一元及多元线性回归分析发现: 1、四区位因素对空间集聚确有正向影响,尤其是资本资源和基础设施。

Chapter three, points out the shortcomings of the present corporate governance of our state-owned enterprises, standing out the irrationality of equity structure is the key reason of the present bad corporate governance. Through both the canonical analysis and the empirical test on the monopoly stock equity structure, oligopoly stock equity structure, complete competition stock equity structure and competitive stock equity structure, it concludes that competitive stock equity structure is advantageous to corporate governance.

At each of 678 sampling locations in China, 15 ecological factors, including 3 geographical factors( x ) and 12 climatic factors( y ), were selected and subjected to the canonical analysis(CA).

Gower's (1977) method for the canonical analysis of asymmetric matrices (CAA) where the same entities index both the rows and the columns is described.

A cluster analysis was performed to supplement the generated information by the canonical analysis.

The relevant summary of the canonical analysis is given in Table 5.

Infrapopulations have been projected on the first two axes of the canonical analysis.

Two methodes to obtain the canonical transformation matrix are presented in this paper. One is to calculate the matrix from the canonical analysis of the field spectral data of objects; The other derives it directly from the DN values of the targets in digital image. The relationship between the results of canonical analysis of field reflectance data and that of DN values of same objects is discussed also in this paper. At last, two examples using those methods for lithologic discrimination...

Two methodes to obtain the canonical transformation matrix are presented in this paper. One is to calculate the matrix from the canonical analysis of the field spectral data of objects; The other derives it directly from the DN values of the targets in digital image. The relationship between the results of canonical analysis of field reflectance data and that of DN values of same objects is discussed also in this paper. At last, two examples using those methods for lithologic discrimination in remote sensing images are given.

The canonical analysis can deal with two kinds of variables simultaneously and the interrelated coefficient results. Ten plant associations and their soil characteristics of Yellow River delta have been analysed by using this method. The canonical interrelated coefficients are 0.9 and 1. All the canonical variables have been calculated with the whole information of primary data retained. It is shown that there exists a close interrelation between the dynamic changes of vegetation and its soil...

The canonical analysis can deal with two kinds of variables simultaneously and the interrelated coefficient results. Ten plant associations and their soil characteristics of Yellow River delta have been analysed by using this method. The canonical interrelated coefficients are 0.9 and 1. All the canonical variables have been calculated with the whole information of primary data retained. It is shown that there exists a close interrelation between the dynamic changes of vegetation and its soil characters. The dynamic change of soil water and salt is the restrictive factor of vegetation succession. The canonical variable ordination shows directly the succession relationship between plant associations. This method offers a great help to the study of plant community distribution and vegetation succession.

At each of 678 sampling locations in China, 15 ecological factors, including 3 geographical factors( x ) and 12 climatic factors( y ), were selected and subjected to the canonical analysis(CA). The coordinate values of the first canonical variable were chosen and subjected further to principle component analysis(PCA) if the canonical correlation was significant at 0.01 level. The sample locations′ coordinate values of the first principle component were taken as comprehensive...

At each of 678 sampling locations in China, 15 ecological factors, including 3 geographical factors( x ) and 12 climatic factors( y ), were selected and subjected to the canonical analysis(CA). The coordinate values of the first canonical variable were chosen and subjected further to principle component analysis(PCA) if the canonical correlation was significant at 0.01 level. The sample locations′ coordinate values of the first principle component were taken as comprehensive ecological gradient axis, namely EGA(CA 1 PC 1), according to the contribution of principle components. The isogram of the ecological gradient axis(EGA) was drawn out. Meanwhile the multinomial regression of the sample locations′ EGA values against their correspondent longitude( x 1) and latitude( x 2) and altitude ( x 3) in built. Then the ecological gradient map of China was plotted by computer according to the isogram and the multinomial regression equation. The forest tree regionalization was quantitatively conducted on the basis of the map and the result of regionalization was further fitted and corrected in accordance with the principles of regionalization. Finally, the whole county was divided into 10 forest tree breeding region groups which were further subdevided into 97 breeding regions, 74 sub regions, and 31 forest tree breeding basic regions for arranging the forest breeding trials.