| 地质数据回归分析的自变量迭代变换 |
| |
| 引用本文： | 张启锐.地质数据回归分析的自变量迭代变换[J].地质科学,1986,0(4):403-410. |
| |
| 作者姓名： | 张启锐 |
| |
| 作者单位： | 中国科学院地质研究所 |
| |
| 摘 要： | 回归分析是地质数据处理中应用最广的一种统计分析方法,尤其是其中的线性回归模型应用最广泛。模型中的变量,包括因变量和自变量,均被看作是线性的,但实际计算时,这些变量却可以看成是原始观测值的一种非线性的变换。
|
| 关 键 词： | , |
| 收稿时间： | 1984-09-01 |
| 修稿时间： | 1984-09-01; |
THE ITERATIVE TRANSFORMATION OF INDEPENDENT VARIABLES IN REGRESSION ANALYSIS OF GEOLOGICAL DATA |
| |
| Zhang Qirui.THE ITERATIVE TRANSFORMATION OF INDEPENDENT VARIABLES IN REGRESSION ANALYSIS OF GEOLOGICAL DATA[J].Chinese Journal of Geology,1986,0(4):403-410. |
| |
| Authors: | Zhang Qirui |
| |
| Institution: | Institute of Geology, Academia Sinica, Beijing |
| |
| Abstract: | It is well hnown that the independent variables in a regression equation can be some transformed versions of their original ones,though the regression model remains to be linear. For a specific variable there are unlimited ways of transformation,and addition-ally,the particular form of transformation of a variable in a model is usually,if not always,unknown. Therefore,looking for the appropriate format of variables in a regression model becomes an important part of the analysis and takes a lot of time and consideration,yet the resulting model is not granted to be optimal.The iterative transformation method introduced in this paper is proved to be useful in overcoming the above-mentioned difficulties. The computational procedure of the method is simple and illustrated in detail. The convergence is usually reached within two or three iterations. Occasionally,unlimited iteration may occur and two possible ways to avoid it are suggested,e.g. setting a maximum iteration number to the computer program (generally it equals 10) or to adjust the threshold value before the next run of the program.If the dependent and independent variables in a model are statistically independent,such as an example given in the text where the values of Pearson correlation coefficient equals -0.13,the results will be unreasonable,and the computation will soon be aborted due to overflow.Two sets of data together with their ordinary regression analysis results were collected from published materials. The data were reanalyzed by iterative method. In the first example,the results are much better than the ordinary ones,and the second example indicated that results of iterative method is comparable to the ordinary ones,if not better. |
| |
| Keywords: | , |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《地质科学》浏览原始摘要信息 |
| 点击此处可从《地质科学》下载免费的PDF全文 |
|
