三种预测林木生长量方法的比较  

Comparative Study of Three Methods for Predicting the Forest Growth

作　　者：马友平[1] 冯仲科[1] 何友均[2] 

机构地区：[1]北京林业大学测绘与3S技术中心 [2]中国林业科学研究院林业科技信息研究所,北京100091

出　　处：《生物数学学报》2007年第2期336-342,共7页Journal of Biomathematics

基　　金："863"国家高新技术项目(2003AA245030);国家自然科学基金项目(90302014);北京市自然科学基金项目(4041002)

摘　　要：本研究选择了三种方法灰色GM(1,1)模型法、回归法、灰色线性回归组合模型法,经过数据的加载,灰色GM(1,1)为Ⅱ级合格模型,发展系数|a|＜0.3可用于中长期预测;在回归法中选用5个模型进行拟合,以Richards方程的复相关系高,残差平方和最小,被选中;灰色线性回组合模型达到Ⅲ级勉强可用模型.用三个模型分别对32、34龄阶的生长量进行预测,其平均相对精度分别为94.64%、80.68%、92.41%.The Grey-Model （1,1）, regression method and the model combination of Grey- Linear regression were selected to predict the forest growth in this paper. Through data loading, The precision of Grey Model （1,1） comes to second level, its expand modulus ｜a｜ was less than 0.3, so the model could be used as medium-term or long-term forecasting on the forest growth. Among the 5 regression model, Richards equation was selected for its high complex dependent coefficient and the minimum value of residual sum of squares. The precision of Grey-Linear regression comes to third level, so the model could be used except no other ways. The growth of 32, 34 aged forest st, and were predicted by above 3 modes, its mean relative precision is 94.64%, 80.68%, 92.41% respectively.

关 键 词：灰色GM(1 1)模型 灰色线性回归组合模型 回归分析 RICHARDS方程 林分生长量 

分 类 号：S758.52[农业科学—森林经理学] S791.223[农业科学—林学]