基于岭回归的思茅松单木生物量复合模型构建

为进一步提高思茅松单木生物量模型的估计精度,采用与材积兼容的生物量模型构造各分量模型,再筛选出最优模型形式构造了线性和非线性复合模型,通过模型4个评价指标对拟合的简单非线性模型,线性和非线性复合模型进行评价。结果表明:思茅松总生物量线性复合模型系统偏差较小,与数据拟合程度更好,预测精度更高,具有可解释性,并且以该方法构造的生物量模型能进一步提高估计精度。

Establishment of Single-Tree Biomass Complex Model for Pinus kesiya var.langbianensis Based on Ridge Regres-sion

For further improving estimate accuracy of single-tree biomass model of Pinus kesiya var.langbianensis, we used structuring component model to build the compatibility biomass model to volume of timber, linear and nolinear complex model with optimal component model, and evaluated simple nonlinear model, linear and nonlinear model with four evalua-tion index.Total biomass linear composite model system of P.kesiya var.langbianensis was with small error, better fitting, higher prediction accuracy, and it can be interpreted to linear complex model.