| 基于分位数回归的针阔混交林树高与胸径的关系 |
| |
| 引用本文： | 张冬燕,王冬至,李晓,高雨珊,李天宇,陈静.基于分位数回归的针阔混交林树高与胸径的关系[J].浙江农林大学学报,2020,37(3):424-431. |
| |
| 作者姓名： | 张冬燕 王冬至 李晓 高雨珊 李天宇 陈静 |
| |
| 作者单位： | 1.河北农业大学 林学院，河北 保定 0710002.河北农业大学 经济管理学院，河北 保定 071000 |
| |
| 基金项目： | 河北省教育厅科研资助项目(QN2018125)；国家自然科学青年基金资助项目(31700561)；“十三五”国家重点研发计划项目(2016YFD060020303，2017YFD0600403) |
| |
| 摘 要： | 目的 基于包含哑变量的非线性分位数回归方法，构建华北落叶松Larix principis-rupprechtii与白桦Betula platyphylla针阔混交林树高与胸径关系的预测模型，对研究混交林中树种结构及立地生产力预测具有重要意义。 方法 以河北省塞罕坝机械林场华北落叶松与白桦针阔混交林为研究对象，利用83块标准地调查数据，基于哑变量分别采用最小二乘法和非线性分位数回归方法，构建不同树种树高与胸径关系模型。 结果 基于包含哑变量的非线性分位数回归预测模型精度高于最小二乘法，其中利用最小二乘法拟合不同树种模型，其确定系数、平均差及平均绝对误差分别为0.787~0.814、1.581~1.877、2.447~2.654；而利用非线性分位数回归不同树种不同分位点模型，其确定系数、平均差及平均绝对误差分别为0.839~0.921、0.213~1.469、0.561~2.322，经过残差分析确定，当分位点τ=0.7时，不同树种树高与胸径关系预测模型精度较高。 结论 与最小二乘法相比，基于非线性分位数构建的包含哑变量不同树种树高与胸径关系的预测模型精度更高。图3表3参33
|
| 关 键 词： | 森林生态学 非线性分位数回归 树高-胸径 哑变量 混交林 |
| 收稿时间： | 2019-08-05 |
Relationship between height and diameter at breast height(DBH) in mixed coniferous and broadleaved forest based on quantile regression |
| |
| ZHANG Dongyan,WANG Dongzhi,LI Xiao,GAO Yushan,LI Tianyu,CHEN Jing.Relationship between height and diameter at breast height(DBH) in mixed coniferous and broadleaved forest based on quantile regression[J].Journal of Zhejiang A&F University,2020,37(3):424-431. |
| |
| Authors: | ZHANG Dongyan WANG Dongzhi LI Xiao GAO Yushan LI Tianyu CHEN Jing |
| |
| Institution: | 1.College of Forestry, Hebei Agricultural University, Baoding 071000, Hebei, China2.College of Economic and Management, Hebei Agricultural University, Baoding 071000, Hebei, China |
| |
| Abstract: | Objective With the employment of nonlinear quantile regression method using dummy variables, the current study is aimed to establish a prediction model for the relationship between height and diameter at breast height (DBH) in Larix principis-rupprechtii and Betula platyphylla mixed forest so as to better predict the tree structure and site productivity of mixed forests. Method Taking L. principis-rupprechtii and B. platyphylla mixed forest of Saihanba Mechanised Tree farm in Hebei Province as the research object, with 83 pieces of standard land survey data used and dummy variables created, this paper adopted the least square method and nonlinear quantile regression method respectively in the construction of the relationship model of tree height and DBH of different species. Result The accuracy of the nonlinear quantile regression prediction model based on dummy variables was higher than that of the one constructed using the least square method. Specifically, when the least square method was used to fit the tree height and DBH relationship model of different tree species, the determination coefficient, average difference and average absolute error of different tree species models were within the range of 0.787?0.814, 1.581?1.877 and 2.447?2.654 respectively. When the nonlinear quantile regression method was used, the coefficient, average deviation, and average absolute error were within the range of 0.839?0.921, 0.213?1.469, 0.561?2.322. In accordance with the residual analysis, when the quantiles of τ is 0.7, the relationship model of tree height and DBH of different species demonstrated a higher accuracy. Conclusion To sum up, compared with the one constructed employing the least square method, the prediction model of tree height and DBH relationship of different tree species adopting the nonlinear quantile regression method has higher prediction accuracy. Ch, 3 fig. 3 tab. 33 ref.] |
| |
| Keywords: | |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《浙江农林大学学报》浏览原始摘要信息 |
| 点击此处可从《浙江农林大学学报》下载免费的PDF全文 |
|
