论文部分内容阅读
该研究在野外获得了鞭角华扁叶蜂虫卵数与生境因子海拔、胸径、树高、海拔、冠幅坡向的数据;内业利用Matlab7软件分别建立了鞭角华扁叶蜂虫卵数的全变量模型、逐步回归模型和主成分模型,三模型的复相关系数R~2分别为0.9204,0.91841和0.9289,均方误差RMSE分别为322.4069,331.7300和310.9550,从两个系数看,得出一致结论以主成分法拟合最佳.逐步回归模型只采用了两个因子即树高和冠幅,主成分法采用了前三个主成分量,其信息量达93.6359%,可代表原特征变量的大部分信息.
In this study, we obtained the data of egg number and habitat height, DBH, tree height, altitude and crown aspect in the field. The full variable model, the stepwise regression model and the principal component model, the complex correlation coefficient R ~ 2 of the three models were 0.9204,0.91841 and 0.9289 respectively, and the mean square error RMSE was 322.4069, 331.7300 and 310.9550, respectively. Judging from the two coefficients, The concordant conclusion is best fitted by the principal component method.The stepwise regression model uses only two factors, tree height and crown width, the principal component method uses the first three principal components, the amount of information is 93.6359%, can represent the original Much of the information about feature variables.