论文部分内容阅读
利用收集的福建省安溪县白濑国有林场48组福建柏树高、胸径数据,按照常规方法建立了福建柏树高—胸径一元线性模型;在此基础上,引入Box-Cox变换,对该模型进行改进。经比较变换前后的因量残差的相伴概率值可知,变换前的相伴概率值P1=0.329,变换后的相伴概率值P2=0.344,较前者提高0.015;比较变换前后的模型估测精度可知,变换前福建柏树高—胸径一元线性模型的估测精度为A1=92.08%,变换后的模型估测精度A2=96.78%,较前者提高了4.70个百分点,福建柏树高—胸径一元线性模型得到改进。
Based on the collected data of height and DBH of 48 sets of cypress trees in Baishui National Forest Farm of Anxi County, Fujian Province, a linear model of height-DBH of Cypress was established according to the conventional method. On this basis, Box-Cox transformation was introduced to improve the model . Comparing the concomitant probability values of the residuals before and after the transformation, we can see that the pre-transformation concomitant probability value P1 = 0.329 and the transformed concomitant probability value P2 = 0.344, which is 0.015 higher than the former. Compared with the former, The accuracy of unary linear model of cypress tree height-diameter before translation was A1 = 92.08%, the estimated accuracy of transformed model A2 = 96.78%, which was 4.70% higher than that of the former, and the linear regression model was improved .