Data and program: Comparison between Machine Learning Models and Conventional Statistical Models in Predicting Global Tree Canopy Height and Crown Radius

The attachment includes three folders:The first folder, Data classification (testing and training), consists of two folders (crown_radius and height), the first crown_radius folder It contains excel data of three plant functional types (PFTs) - temperate needleleaf trees (MN), temperate broadleaf trees (MB) and tropical broadleaf trees (TB), these three excel data all contain 19 soil factors data, 22 climate factors data and information such as crown_radius_m, mask, stem_diameter_cm, etc. The information in the second height folder is similar, and it corresponds to Table 1.Data summary and Figure 3 for each PFT in the article; The second folder, Feather importance, contains two excel spreadsheets (crown_radius-FI and height-FI), the first excel spreadsheet of crown_radius-FI Feather importance containing three plant functional types (PFTs) is temperate needleleaf trees (MN), temperate broadleaf trees (MB), and tropical broadleaf trees (TB); The excel table information of the second height-FI is similar, and its information corresponds to Figure 5 and Figure S3 in the article; The third folder "program" contains two packages (make_model1 and make_model2) and a calling program "Source program". Among them, the make_model1 package is mainly used to obtain the best parameters for selecting the model; The make_model2 package is based on the selection of the make_model1 package to further analyze the specific FI values of the factors in the best model. The Source program is used to make specific calls to the package according to the requirements.附件包含三个文件夹：第一个文件夹为数据分类（测试集与训练集），下设冠幅半径（crown_radius）与树高（height）两个子文件夹。其中冠幅半径文件夹内存储了3种植物功能型（Plant Functional Types, PFTs）的Excel数据：温带针叶树（MN）、温带阔叶树（MB）与热带阔叶树（TB）；每份Excel数据均包含19项土壤因子、22项气候因子，以及冠幅半径（crown_radius_m）、掩膜（mask）、茎干直径（stem_diameter_cm）等相关信息。树高文件夹的内容结构与之类似，其数据对应本文中各植物功能型的表1数据汇总与图3。 第二个文件夹为特征重要性（Feature Importance），内含两份Excel表格：冠幅半径-特征重要性（crown_radius-FI）与树高-特征重要性（height-FI）。其中冠幅半径-FI表格包含温带针叶树（MN）、温带阔叶树（MB）与热带阔叶树（TB）这3种植物功能型的特征重要性数据；树高-FI表格的信息结构与之类似，其数据对应本文中的图5与补充图S3。 第三个文件夹为程序（program），内含两个工具包（make_model1与make_model2）与一个调用程序「源程序（Source program）」。其中，make_model1工具包主要用于获取模型选择所需的最优参数；make_model2工具包基于make_model1选定的最优模型，进一步分析各因子在最优模型中的具体特征重要性数值；源程序则用于根据需求对工具包进行针对性调用。