stacking 集成学习模型在森林蓄积量估测中的应用

Application of stacking Ensemble Learning Model in Forest volume Estimation

摘要:

为 了探索 sentinel-2遥感数据在估测森林蓄积量中的适用性,以及开发一种提高蓄积量估测精度的集成学习算 法,选择江西省兴国县为研究区 , 以 sentinel-2为遥感数据源,利用 Boruta 算法进行特征选择后开发了 一种 stacking 集成 学习模型,并且与 MLR、KNN、sVR和 RF 四种基础模型进行对比。 结果表明,相比于 MLR模型,机器学习模型具有更强 的蓄积量估测能力,利用机器学习模型估计森林蓄积量的 RMsE 降低 了 18. 02~22. 50 m3 ● hm- 2 , rRMsE 降低 了 9. 01%~11. 25%。 并且相比于基础模型,利用 stacking 算法将 4 种模型进行集成后,模型的均方根误差进一步降低 了 11. 95~7. 47m3 ● hm-2 ,说明 stacking集成学习算法可以有效的提高森林蓄积量的估测性能。

Abstract:

In order to explore the applicability oFSentinel-2remote sensing data in estimating Forest stocKand to develop an integrated learning algorithm to improve the accuracy oFstocK estimation, xingguo county, Jiangxi province was se- lected study area, uses Sentinel-2 as the remote sensing data source , develops a stacKing integrated learning model using Boruta algorithm For Feature selection, and compares it with Four basic models, namely MLR, KNN, SVR and RF. The results showed that compared with the MLR, KNN, SVR and RF models, the stacKing integrated learning model was more eFFicient than the MLR model. The machine learning model has stronger accumulation estimation ability than the MLR model, and the RMSE oFForest accumulation estimation using the machine learning model was reduced by 18. 02~ 22. 50m3 ● hm—2 and the rRMSEwas reduced by 9. 01~11. 25percentage points . In addition, the RMSE oFthe model was Further reduced by 11. 95~7. 47m3 ● hm—2 aFter integrating the Four models using the StacKing algorithm compared with the base model, indicating that the StacKing integrated learning algorithm can eFFectively improve the estimation perForm- ance oFForest stocK.